DRAFT Guidance for Ozone and Fine Particulate Matter Permit Modeling

UNITED STATES ENVIRONMENTAL PROTECTION AGENCY RESEARCH TRIANGLE PARK, NC 27711

n:e 1 o 2020

OFFICE OF AIR QUALITY PLANNING MEMORANDUM AND STANDARDS

SUBJECT: DRAFT Guidance for Ozone and Fine Particulate Matter Permit Modeling

FROM: Richard A. Wayland, Division Director��A. w7a....o,(_ Air Quality Assessment Division

TO: Regional Air Division Directors, Regions 1 - 10

The Environmental Protection Agency (EPA) is providing the attached DRAFT Guidance for Ozone and Fine Particulate Matter Permit Modeling to the state, local, and tribal air agencies, as well as the public, for consideration, review and comment. This guidance document reflects the EPA's recommendations for how a stationary source seeking a Prevention of Significant Deterioration (PSD) permit may demonstrate that it will not cause or contribute to a violation of the National Ambient Air Quality Standards (NAAQS) for ozone (03) and fine particulate matter (PM2.s) and PSD increments for PM2.s, as required under Section 165(a)(3) of the Clean Air Act (CAA) and 40 CFR sections 5 l.l 66(k) and 52.21 (k).

This document does not substitute for provisions or regulations of the CAA, nor is it a regulation itself. As the term "guidance" suggests, it provides recommendations on how to implement the modeling requirements of a PSD compliance demonstration. Thus, it does not impose binding, enforceable requirements on any party, nor does it assure that the EPA will approve all instances of its application, as the guidance may not apply to a particular situation based upon the circumstances. Final decisions by the EPA regarding a particular PSD compliance demonstration will only be made based on the statute and applicable regulations, and will only be made followinga final submission by air agencies and after notice and opportunity for public review and comment.

BACKGROUND

The EPA is providing this DRAFT Guidance for Ozone and Fine Particulate Matter Permit Modeling to fulfill an outstanding need foradditional guidance on demonstrating compliance with the NAAQS for 03 and PM2.s and PSD increments for PM2.s. Because of the complex chemistry of secondary formationof 03 and PM2.s, the EPA's judgment in the past was that it was not technically sound to specify with "reasonable particularity" air quality models that must be used to assess the impacts of a single source on 03 and secondary PM2.s concentrations. Instead, the EPA employed a case-by-case process fordetermining analytical techniques that

InternetAddress (URL) • http://www.epa.gov Recycled/Recyclable • Printed wHh VegelableOIi Based Inks on Recycled Paper{ Minimum 25% Postconsumer) should be used for these secondary pollutants. However, as discussed in the preamble of the 2017 revisions to the EPA’s Guideline on Air Quality Models1:

“…the EPA has determined that advances in chemical transport modeling science indicate it is now reasonable to provide more specific, generally-applicable guidance that identifies particular models or analytical techniques that may be used under specific circumstances for assessing the impacts of an individual or single source on ozone and secondary PM2.5. For assessing secondary pollutant impacts from single sources, the degree of complexity required to appropriately assess potential impacts varies depending on the nature of the source, its emissions, and the background environment. In order to provide the user community flexibility in estimating single-source secondary pollutant impacts that allows for different approaches to credibly address these different areas, the EPA proposed a two-tiered demonstration approach for addressing single-source impacts on ozone and secondary PM2.5.”

This recommended two-tiered demonstration approach was promulgated as part of the 2017 Guideline revisions.

This draft guidance provides an update to the previous Guidance for PM2.5 Permit Modeling2 to reflect the 2017 revisions to the Guideline and incorporate appropriate sections for O3. As experience is gained with these types of PSD compliance demonstrations, the EPA expects to update this and related guidance and provide further specificity on procedures for assessing the impacts of a single source on O3 and secondary PM2.5 concentrations.

REVIEW AND COMMENT

The EPA is requesting that comments on the draft guidance be provided by Friday, March 27, 2020. This allows at least 45 days for consideration, review, and comment on the material presented in the draft guidance. Comments should be electronically submitted to Mr. George Bridgers of the EPA’s Air Quality Modeling Group at [email protected].

Following the close of the comment period, the EPA will take into consideration all the feedback and comments submitted and will further engage with the regulatory air quality modeling community at the 2020 Regional, State, and Local Modelers’ Workshop currently scheduled for May 5-7, 2020, at the Minneapolis Central Library in Minneapolis, MN. This workshop will allow for an open dialogue on further clarifications, potential amendments, and considerations for additions to the final guidance documentation to be released later this year.

1 Guideline on Air Quality Models. 40 CFR part 51, Appendix W (82 FR 5182, Jan. 17, 2017). https://www3.epa.gov/ttn/scram/guidance/guide/appw_17.pdf. Also know as the “2017 Guideline.” 2 Guidance for PM2.5 Modeling. May 20, 2014. Publication No. EPA-454/B-14-001. Office of Air Quality Planning and Standards, Research Triangle Park, NC. https://www3.epa.gov/ttn/scram/guidance/guide/Guidance_for_PM25_Permit_Modeling.pdf.

The EPA will also conduct a webinar providing an overview of the DRAFT Guidance for Ozone and Fine Particulate Matter Permit Modeling allowing for an open exchange on the guidance documentation on Thursday, March 12th at 3pm EDT. Additional information on how to connect to the webinar is posted on the EPA’s SCRAM website, https://www.epa.gov/scram, under the Recent Additions section and will be shared with the regulatory air quality modeling community through typical email distributions.

For convenience, the draft guidance document is available electronically on the EPA’s SCRAM website at: https://www3.epa.gov/ttn/scram/guidance/guide/Draft_Guidance_for_O3_PM25_Permit_Modeling.pdf.

If there are any questions regarding the draft guidance, please contact George Bridgers of EPA’s Air Quality Modeling Group at (919) 541-5563 or [email protected].

cc: Air Program Managers, EPA Regions 1 – 10 Peter Tsirigotis, OAQPS Mike Koerber, OAQPS Scott Mathias, OAQPS, AQPD Raj Rao, OAQPS, AQPD Brian Doster, OGC Stephanie Hogan, OGC Mark Kataoka, OGC

Attachment

DRAFT Guidance for Ozone and Fine Particulate Matter Permit Modeling

EPA-457/P-20-002 February 2020

DRAFT Guidance for Ozone and Fine Particulate Matter Permit Modeling

U.S. Environmental Protection Agency Office of Air Quality Planning and Standards Air Quality Policy Division Research Triangle Park, NC

This Page Intentionally Left Blank

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

TABLE OF CONTENTS

I. Introduction ...... 1 II. Guidance Overview ...... 7

II.1 Significant Emissions Rates for O3 and PM2.5 ...... 10

II.2 PSD Pollutant Applicability for O3 and PM2.5 ...... 11

II.3 Significant Impact Levels for O3 and PM2.5 ...... 12 II.4 Source Impact Analysis ...... 14 II.5 Cumulative Impact Analysis ...... 15

II.5.1 O3 and PM2.5 NAAQS Compliance ...... 15

II.5.2 PM2.5 PSD Increments Compliance ...... 16

III. PSD Compliance Demonstrations for the O3 and PM2.5 NAAQS: Source Impact Analysis ...... 19

III.1 O3 NAAQS ...... 19

III.2 PM2.5 NAAQS ...... 20

III.3 Assessing Primary PM2.5 Impacts ...... 23

III.4 Assessing O3 and Secondary PM2.5 Impacts ...... 24 III.4.1 Conceptual Model ...... 24 III.4.2 Tier 1 Assessment Approach ...... 26 III.4.3 Tier 2 Assessment Approach ...... 31 III.5 Comparison to the SIL ...... 35

III.5.1 SIL Comparison for O3 ...... 35

III.5.2 SIL Comparison for PM2.5 ...... 36

IV. PSD Compliance Demonstrations for the O3 and PM2.5 NAAQS: Cumulative Impact Analysis ...... 41 IV.1 Modeling Inventory ...... 43 IV.2 Monitored Background ...... 45 IV.3 Comparison to the NAAQS ...... 46 IV.4 Determining Whether Proposed Source Causes or Contributes to Modeled Violations ...... 54

V. PSD Compliance Demonstration for the PM2.5 Increments ...... 57 V.1 Overview of the PSD Increment System ...... 57 V.1.1 PSD Increments and Baseline Concentration ...... 57 V.1.2 PSD Baseline Area and Key Baseline Dates ...... 59 V.1.3 PSD Increment Expansion ...... 62

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

V.2 PSD PM2.5 Increments...... 63

V.3 PSD Compliance Demonstration for the PM2.5 Increments ...... 65

V.3.1 PM2.5 Increments: Source Impact Analysis ...... 65

V.3.2 PM2.5 Increments: Cumulative Analysis ...... 69

V.3.2.1 Assessing Primary PM2.5 Impacts ...... 70

V.3.2.2 Assessing Secondary PM2.5 Impacts ...... 71 V.4. Determining Whether a Proposed Source Will Cause or Contribute to an Increment Violation ...... 72 VI. References ...... 75

Appendix A: Draft Conceptual Description of O3 and PM2.5 Concentrations in the U.S...... A-1

Appendix B: General Guidance on Use of Dispersion Models for Estimating Primary PM2.5 Concentrations ...... B-1

Appendix C: Example of a Tier 1 Demonstration of the Potential for O3 and Secondary PM2.5 Formation...... C-1 Appendix D: Example of the background monitoring data calculations for a Second Level 24-hour modeling analysis ...... D-1

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

I. Introduction

The U.S. Environmental Protection Agency (EPA) is providing this “Guidance for Ozone and Fine Particulate Matter Permit Modeling” to fulfill a need for additional guidance on demonstrating compliance with the ozone (O3) and fine particulate matter (PM2.5) National

Ambient Air Quality Standards (NAAQS) and the Prevention of Significant Deterioration (PSD) increments for PM2.5. Because of the complex chemistry of secondary formation of O3 and

PM2.5, the EPA's judgment in the past was that it was not technically sound to specify with

“reasonable particularity” air quality models that must be used to assess the impacts of a single

source on O3 and secondary PM2.5 concentrations. Instead, the EPA employed a case-by-case

process for determining analytical techniques that should be used for these secondary pollutants.

Under the former process, EPA recommended that the “[c]hoice of methods used to assess the

impact of an individual source depends on the nature of the source and its emissions. Thus,

model users should consult with the Regional Office to determine the most suitable approach on

a case-by-case basis” (2005 Guideline on Air Quality Models, U.S. EPA, 2005; hereafter referred

to as 2005 Guideline; sections 5.2.1.c and 5.2.2.1.c). As such, under the 2005 Guideline, the

appropriate methods for assessing O3 and secondary PM2.5 impacts were determined as part of

the normal consultation process with the appropriate permitting authority.

On January 4, 2012, the EPA granted a petition submitted on behalf of the Sierra Club on

July 28, 2010 (U.S. EPA, 2012), which requested that the EPA initiate rulemaking regarding the establishment of air quality models for O3 and PM2.5 for use by all major sources applying for a

PSD permit. In granting that petition, the EPA committed to engage in rulemaking to evaluate

whether updates to the 2005 Guideline were warranted and, as appropriate, incorporate new

analytical techniques or models for O3 and secondarily formed PM2.5. As discussed in the

1 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

preamble of the 2017 revisions to the EPA’s Guideline on Air Quality Models (U.S. EPA, 2017a;

hereafter referred to as 2017 Guideline), “the EPA has determined that advances in chemical transport modeling science indicate it is now reasonable to provide more specific, generally-

applicable guidance that identifies particular models or analytical techniques that may be used

under specific circumstances for assessing the impacts of an individual or single source on ozone

and secondary PM2.5. For assessing secondary pollutant impacts from single sources, the degree

of complexity required to appropriately assess potential impacts varies depending on the nature

of the source, its emissions, and the background environment. In order to provide the user

community flexibility in estimating single-source secondary pollutant impacts that allows for different approaches to credibly address these different areas, the EPA proposed a two-tiered demonstration approach for addressing single-source impacts on ozone and secondary PM2.5.”

This recommended two-tiered demonstration approach was promulgated as part of the 2017

Guideline revisions.

As presented in section 5.2 of the 2017 Guideline, the first tier involves use of technically credible relationships between precursor emissions and a source’s impacts. Such information may be published in the peer-reviewed literature; developed from modeling that was previously conducted for an area by a source, a governmental agency, or some other entity that is deemed sufficient; or generated by a peer-reviewed reduced form model. To assist permitting authorities, the EPA released the “Guidance on the Development of Modeled Emission Rates for Precursors

(MERPs) as a Tier 1 Demonstration Tool for Ozone and PM2.5 under the PSD Permitting

Program” (U.S. EPA, 2019a; hereafter referred to as MERPs Guidance) that provides a framework to develop MERPs for consideration and use as a Tier 1 demonstration tool, as described in the preamble of the 2017 Guideline.

2 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

The second tier, also presented in section 5.2 of the 2017 Guideline, involves application

of more sophisticated case-specific chemical transport models (CTMs), e.g., photochemical grid

models, to be determined in consultation with the EPA Regional Offices. The EPA provided

guidance to permitting authorities on procedures for applying CTMs in the “Guidance on the Use

of Models for Assessing the Impacts of Emissions from Single Sources on the Secondarily

Formed Pollutants: Ozone and PM2.5” (U.S. EPA, 2016a; hereafter Single-source Modeling

Guidance). The Single-source Modeling Guidance is intended to inform that second tier approach by providing appropriate technical methods to assess O3 and secondary PM2.5 impacts

associated with the precursor emissions from the new or modifying source. The appropriate tier

for a given application should be selected in consultation with the appropriate permitting

authority and be consistent with EPA guidance.

This guidance provides an update to the previous “Guidance for PM2.5 Permit Modeling”

(U.S. EPA, 2014a) to reflect the 2017 revisions to the Guideline and incorporate appropriate sections for O3. As experience is gained with these types of PSD compliance demonstrations, the

EPA expects to update this and related guidance and provide further specificity on procedures for

assessing the impacts of a single source on O3 and secondary PM2.5 concentrations.

This guidance document is organized in three primary areas:

1. Guidance Overview – Section II provides a general overview of the steps that a

permit applicant would take under the PSD program for demonstrating

compliance with the O3 NAAQS and/or the PM2.5 NAAQS and PSD

increments.

2. PSD Compliance Demonstrations for the O3 and PM2.5 NAAQS – Sections III and

IV provide a detailed framework for conducting a source impact analysis and

3 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

a cumulative impact analysis, respectively, to appropriately address O3 and

1 PM2.5 impacts from the proposed source in determining whether it may cause

or contribute to a NAAQS violation.

3. PSD Compliance Demonstrations for PM2.5 Increments – Section V provides a

detailed discussion of the assessment of primary and secondary PM2.5 impacts

of a new or modifying source with respect to the PM2.5 increments.

This document recommends procedures for permit applicants and permitting authorities

to follow to show that they have satisfied some of the criteria for obtaining or issuing a permit

under applicable PSD regulations. This document is not a rule or regulation, and the guidance it

contains may not apply to a particular situation based upon the individual facts and

circumstances. This guidance does not change or substitute for any law, regulation, or any other

legally binding requirement, may refer to regulatory provisions without repeating them in their

entirety, and is not legally enforceable. The use of non-mandatory language such as “guidance,”

“recommend,” “may,” “should,” and “can,” is intended to describe EPA policies and

recommendations. Mandatory terminology such as “must” and “required” are intended to

describe requirements under the terms of the CAA and EPA regulations, but this document does not establish or alter any legally binding requirements in and of itself.

This guidance does not create any rights or obligations enforceable by any party or impose binding, enforceable requirements on any PSD permit applicant, PSD permitting authority, EPA, or any other person. Since each permitting action will be considered on a case-

by-case basis, this document does not limit or restrict any particular justifiable approach that

1 The term “proposed source” is used throughout this guidance document and should be taken to mean the “proposed source or modification” to which the compliance demonstration is being assessed.

4 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

permit applicants and permitting authorities may take to conduct the required compliance

demonstrations. Each individual decision to issue a PSD permit must be supported by a record

sufficient to demonstrate that the proposed construction and operation of a stationary source will

not cause or contribute to a violation of the applicable NAAQS and PSD increments. While this

document illustrates a particular approach that the EPA considers appropriate and acceptable as a

general matter, permit applicants and permitting authorities should examine all relevant

information regarding air quality in the area that may be affected by a proposed new or modified

source and evaluate whether alternative or additional analysis may be necessary in a given case

to demonstrate that the regulatory criteria for a PSD air quality analysis are satisfied. This document does not represent a conclusion or judgment by EPA that the technical approaches recommended in this document will be sufficient to make a successful compliance demonstration in every permit application or circumstance.

Permitting authorities retain the discretion to address particular issues discussed in this document in a different manner than the EPA recommends so long as the approach is adequately justified, supported by the permitting record and relevant technical literature, and consistent with the applicable requirements in the CAA and implementing regulations, including the terms of an approved State Implementation Plan (SIP). Furthermore, this guidance is not a final agency action and does not determine applicable legal requirements or the approvability of any particular permit application. To improve the quality of this guidance, the EPA is soliciting public comment and will consider the comments received.

The EPA Regional Offices may seek clarification from the EPA’s Office of Air Quality

Planning and Standards (OAQPS) on issues and areas of concern in a modeling protocol or PSD compliance demonstration. Through these interactions and subsequent resolutions of specific

5 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

issues, clarifications of preferred modeling procedures can become additional EPA guidance.

This can happen in several ways: 1) the preferred procedures are published as regulations or guidelines; 2) the preferred procedures are formally transmitted as guidance to the Air Division

Directors in the EPA Regional Offices; 3) the preferred procedures are formally transmitted as

guidance to the EPA Regional Office modeling contacts; or 4) the preferred procedures are relied

upon in decisions by the EPA’s Model Clearinghouse that establish national precedent that the

approach is technically sound. The Model Clearinghouse is the EPA focal point for the review of

the technical adequacy of pollutant modeling to satisfy regulatory criteria and other NAAQS

compliance demonstration techniques. Model Clearinghouse memoranda involving interpretation

of modeling guidance for specific applications, as well as other clarification memoranda

addressing modeling more generally, are available at the Support Center for Regulatory

Atmospheric Modeling (SCRAM) website at: https://www.epa.gov/scram/air-quality-model-

clearinghouse.

6 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

II. Guidance Overview

This guidance is appropriate for proposed new or modifying sources locating or located in an area classified as attainment or unclassifiable for O3 and/or PM2.5. It is intended to provide

recommendations on how to conduct compliance demonstrations for the O3 NAAQS and the

PM2.5 NAAQS and PSD increments under the PSD program following the progressive steps

shown in Figure II-1 (for O3 and PM2.5 NAAQS) and Figure II-2 (for PM2.5 increments). Since

each permitting action is considered on a case-by-case basis, this guidance does not limit or restrict any particular justifiable approach that permit applicants and permitting authorities may take to conduct the required compliance demonstrations. Prospective permit applicants should recognize the importance of the consultation process with the appropriate permitting authority.

This process will help identify the most appropriate analytical techniques to be used for conducting a compliance demonstration for the O3 NAAQS and the PM2.5 NAAQS and PSD

increments.

The EPA has historically supported the use of screening tools to help facilitate the

implementation of the PSD program and streamline the permitting process in circumstances

where proposed construction is projected to have an insignificant impact on air quality. These

screening tools include significant emission rates (SERs) and significant impact levels (SILs).

The use of these screening tools at each progressive step on the left side (attainment or

unclassifiable areas) of Figure II-1 and Figure II-2 are described in more detail throughout

Section II.

7 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

Figure II-1. Overview of O3 and PM2.5 NAAQS Compliance Demonstration for New or Modifying Sources under NSR/PSD Programs

New or Modified Source

N Y Attainment or Unclassifiable Nonattainment Nonattainment Area Area Area?

Analysis of ambient air quality Direct or interpollutant offsets Source Emissions N* impacts is NOT required for the Reference the O3 SIP Implementation Greater Than or final rule (83 FR 62998) and the particular pollutant Equal to SER(s)?* PM2.5 NSR Implementation final rule Reference Section III.1 and III.2 (73 FR 28321)

Source Impact Analysis

Source Impact N Compliance demonstration may Greater Than or be adequate Equal to SIL? Reference Section III.5

Cumulative Impact Analysis

Projected N Compliance demonstration may NAAQS be adequate Violations? Reference Section IV.3

Source Impact Greater N Compliance demonstration may Than or Equal to SIL at Projected be adequate Violations? Please reference Section IV.4

Y Compliance demonstration is NOT adequate

* Any emissions rate or any net emissions increase associated with a major stationary source or major modification, which would construct within 10 kilometers of a Class I area, and have an impact on such area equal to or greater than 1 µg/m 3, (24-hour average) is considered significantand should proceed with an appropriate air quality assessment. See 40 CFR 52.21(b)(23)(iii).

8 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

Figure II-2. Overview of PM2.5 PSD Increments Compliance Demonstration for New or Modifying Sources under NSR/PSD Programs

New or Modified Source

N Y Attainment or Unclassifiable Nonattainment Nonattainment Area Area Area?

Increment does not apply Analysis of ambient air quality Source Emissions N* impacts is NOT required for the Greater Than or particular pollutant Equal to SER(s)?* Reference Section V.4.1

First PSD Y Major Source N Construction Since Application in Area Major Source Baseline After Trigger Date? Date?

N Y

Source Impact Analysis

Source Impact N N Compliance demonstration may Source Impact Greater Than or be adequate Above Increment? Equal to SIL? Reference Section V.3.1

Y Y Compliance demonstration is NOT adequate

Cumulative Impact Analysis Compliance demonstration may be adequate Projected N Compliance demonstration may Reference Section V.3.1 Increment be adequate Violations? Reference Section V.3.2

Source Impact Greater N Compliance demonstration may T han or Equal t o SIL at Projected be adequate Violations? Refernce Section V.4

Y Compliance demonstration is NOT adequate

9 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

II.1 Significant Emissions Rates for O3 and PM2.5

O3 and PM2.5 are “regulated NSR pollutant[s]” as that term is defined in the PSD

2 regulations. Pursuant to that definition, ambient concentrations of O3 are generally addressed

through the regulation of its two precursors, nitrogen oxides (NOX) and volatile organic

compounds (VOC), while ambient concentrations of PM2.5 are generally addressed through the

3 regulation of direct PM2.5 and its precursors NOX and sulfur dioxide (SO2). “Significant,” with

respect to O3 and PM2.5, is defined in EPA regulations at 40 CFR 52.21(b)(23) in reference to a

source’s potential to emit (or in the case of a modification, the emissions increase4 and net

emissions increase) either direct emissions of the pollutant or emissions of a precursor pollutant.

The regulations state that an increase in emissions of either O3 precursor (NOX or VOC) is

significant if the increase of the particular precursor equals or exceeds 40 tons per year (tpy). For

direct emissions of PM2.5, the significance level is 10 tpy; for PM2.5 precursor emissions, the

5 significance level is 40 tpy for SO2 and 40 tpy for NOX.

2 40 CFR 52.21(b)(50). 3 See 73 FR at 28333. The EPA’s PSD regulations do not presumptively require VOC to be treated as precursors to PM2.5 in the PSD program. However, a state or the EPA may demonstrate that VOC emissions in a specific area are a significant contributor to that area’s ambient PM2.5 concentrations and, thus, should be treated as a regulated NSR pollutant subject to the PSD permitting requirements. 40 CFR 52.21(b)(50)(i)(b)(4). 4 While section 52.21(b)(23) explicitly defines “significant” for purposes of a net emissions increase or potential to emit, section 52.21(b)(40) defines “significant emissions increase” by reference to the definition of “significant” found in paragraph (b)(23). 5 A significance rate for VOC as a PM2.5 precursor is not defined in the PSD regulations. However, the EPA’s final rulemaking action promulgating regulations for implementing the PSD permitting requirements for PM2.5 and its precursors indicated that any state required to regulate VOC emissions as a PM2.5 precursor “would be required to adopt the 40-tpy significant emissions rate unless it demonstrates that a more stringent significant emissions rate (lower rate) is more appropriate.” 73 FR at 28333.

10 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

II.2 PSD Pollutant Applicability for O3 and PM2.5

The EPA’s PSD regulations apply specific permitting requirements (e.g., Best Available

Control Technology (BACT) and air quality analysis) to regulated New Source Review (NSR)

pollutants that would be emitted in a significant amount by a proposed new or modified major

stationary source.6 For a new major stationary source, PSD permitting requirements apply to any

regulated NSR pollutant for which the source would have the potential to emit a significant

amount. For a modification at an existing major stationary source, PSD permitting requirements

apply to any regulated NSR pollutant for which the modification would result in a significant

emissions increase and a significant net emissions increase (i.e., a “major modification”) of that pollutant.

The provisions at 40 CFR 52.21(m)(1) and (k)(1) comprise the preconstruction air quality

analysis requirements of the PSD program and apply to each regulated NSR pollutant that the source or modification would emit in a significant amount. Paragraph (m)(1) provides that any

PSD permit application shall contain an analysis of ambient air quality for each such pollutant, and paragraph (k)(1) provides that the owner or operator “shall demonstrate that allowable emission increases from the proposed source or modification . . . would not cause or contribute to air pollution in violation of [any NAAQS or PSD increment].”7 EPA interprets the term

“allowable emission increases” as it is used in paragraph (k)(1) to mean those emission increases

authorized by the PSD permit, so that, consistent with paragraph (m)(1), the requirement applies

to regulated NSR pollutants that would be emitted in a significant amount.

6 See 40 CFR 52.21(a)(2) for applicability procedures for new or modified major stationary sources. 7 In accordance with CAA § 165(e)(2), one purpose of the monitoring requirements contained in 40 CFR 52.21(m) is to provide information relevant to the determination of whether emissions from a proposed source or modification will exceed a NAAQS or PSD increment. Therefore, EPA reads paragraphs (m) and (k) of 40 CFR 52.21 together.

11 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

With respect to the unique nature of the criteria pollutants O3 and PM2.5 emissions of

individual O3 and PM2.5 precursors (i.e., NOX, VOC, SO2, and direct PM2.5 are not summed

when determining a significant emissions increase for either criteria pollutant.8 Only precursors of O3 or PM2.5 that would by themselves be emitted by the source in a significant amount are

included in the air quality analysis.

II.3 Significant Impact Levels for O3 and PM2.5

The EPA has issued guidance recommending that permitting authorities consider the use

of appropriate pollutant-specific concentration levels known as “significant impact levels”

(earlier referred to as SILs) as a compliance demonstration tool for O3 and PM2.5 air quality

assessments on case-by-case basis in PSD permitting actions (U.S. EPA 2018a). The “SILs

Guidance” identified recommended SIL values for the O3 and PM2.5 NAAQS and the PM2.5 PSD

increments and included a policy document, as well as supporting technical and legal analyses,

that EPA and other permitting authorities may use in case-by-case PSD permitting actions. As

explained in the guidance, if a permitting authority chooses to use a recommended SIL value to

support a PSD permitting decision, it should justify the values and their use in the administrative

record for the permitting action and may choose to adopt EPA’s SILs Guidance, including the

supporting technical and legal documents, in doing so.

The EPA’s recommended SIL values from the SILs Guidance for the O3 and PM2.5

NAAQS are presented in Table II-1 and for the PM2.5 PSD increments in Table II-2. It is

important to note that the PM2.5 NAAQS has two averaging periods: 24-hour and annual. There

8 See 57 FR 55620, 55624 (Nov. 25, 1992); 80 FR 65292, 65441 (Oct. 26, 2015); see also 73 FR 28321, 28331 (May 16, 2008).

12 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

are no PSD increments established for O3 and, thus, no O3 increments SIL values. For a full

discussion of the basis and purpose of the recommended O3 and PM2.5 SIL values, see the SILs

Guidance and supporting documents (U.S. EPA 2018a).

Table II-1. EPA Recommended SIL Values for O3 and PM2.5 NAAQS Criteria Pollutant (NAAQS Level) NAAQS SIL Concentration Ozone 8-hour (70 ppb) 1.0 ppb 3 3 PM2.5 24-hour (35 µg/m ) 1.2 µg/m 3 3 3 PM2.5 Annual (12 µg/m or 15 µg/m ) 0.2 µg/m

Table II-2. EPA Recommended SIL Values for PM2.5 PSD Increments PSD Increment SIL Concentration Criteria Pollutant Class I Class II Class III 3 3 3 PM2.5 24-hour 0.27 µg/m 1.2 µg/m 1.2 µg/m 3 3 3 PM2.5 Annual 0.05 µg/m 0.2 µg/m 0.2 µg/m

As explained in the SILs Guidance, SILs are designed to have a role throughout the PSD air quality compliance demonstration. A permitting authority that chooses to use SILs would initially compare the modeled concentrations resulting from the proposed source’s emissions increase to the appropriate SIL. This initial comparison is the “Source Impact Analysis.” Where

the proposed source’s projected impacts on air quality concentrations are found at this first stage

to be greater than or equal to the level of the applicable SIL, the analysis should then proceed to

a second stage, which involves a cumulative assessment of the air quality in the affected area.

The “Cumulative Impact Analysis” considers the combined impact of the proposed source or

modification and other relevant sources in determining whether there would be a violation of any

NAAQS or PSD increment in the affected area and, if so, whether the proposed source or

modification would cause or contribute to such violation based on the applicable SIL.

13 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

II.4 Source Impact Analysis

As described in section 9.2.3 of the 2017 Guideline, the EPA’s recommended procedure

for conducting a PSD air quality assessment is a multi-stage approach. The first step is a source

impact analysis that quantifies the air quality concentration increase expected to result from a

new or modifying source’s significant emissions increase as proposed in the PSD permit

application.9 The source impact analysis is used to assess the potential of a proposed new or

modifying source to cause or contribute to a NAAQS or PSD increment violation.

In a source impact analysis, as illustrated in Figure II-1 and Figure II-2 and further

explained in this guidance, a permitting authority compares the modeled concentrations resulting

from the proposed source’s emissions increase to an appropriate O3 or PM2.5 SIL. If the proposed source’s maximum modeled impacts are found to be below the level of the O3 or PM2.5 SIL at

every modeled receptor, the findings of the source impact analysis may be sufficient to

demonstrate that the source will not cause or contribute to a violation of the O3 NAAQS, PM2.5

NAAQS, or the PM2.5 PSD increment, as necessary to receive a PSD permit. On the other hand,

where the proposed source’s projected impacts on air quality concentrations are estimated to be

greater than or equal to the level of an appropriate O3 or PM2.5 SIL at any modeled receptor, the demonstration should proceed to the next step of conducting a cumulative impact analysis.

9 This is consistent with EPA’s overall approach for the use of screening techniques in air quality modeling. See 40 CFR part 51, Appendix W, sections 2.2 (“Levels of Sophistication of Air Quality Analyses and Models”) and 4.2.1 (“Screening Models and Techniques”). In section 2.2.a, the Guideline observes that “[it] is desirable to begin an air quality analysis by using simplified and conservative methods followed, as appropriate, by more complex and refined methods. The purpose of this approach is to streamline the process and sufficiently address regulatory requirements by eliminating the need of more detailed modeling when it is not necessary in a specific regulatory application. For example, in the context of a PSD permit application, a simplified and conservative analysis may be sufficient where it shows the proposed construction clearly will not cause or contribute to ambient concentrations in excess of either the NAAQS or the PSD increments.”

14 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

II.5 Cumulative Impact Analysis

This section provides an overview of cumulative impact analyses for O3 and PM2.5

NAAQS, as well as, PSD increments compliance. The cumulative impact analysis is illustrated in Figure II-1 and Figure II-2 and further explained in this guidance.

II.5.1 O3 and PM2.5 NAAQS Compliance

For either O3 or PM2.5, where the source impact analysis described in Section II.4 is

insufficient to show that a proposed PSD source will not cause or contribute to a violation of the

respective NAAQS, a cumulative impact analysis is then necessary to make the required

NAAQS demonstration, as described in section 9.2.3 of the 2017 Guideline. A cumulative impact analysis should account for the combined impacts of the following:

1. Direct and/or precursor emissions that the new or modifying source would emit in

significant amounts;10

2. Direct emissions from nearby sources (for primary PM2.5 only), as appropriate; and

3. Monitored background levels that account for secondary impacts from regional

background sources, secondary impacts from precursor emissions from nearby

sources, and, in the case of primary PM2.5, PM2.5 impacts from direct emissions from

background sources, nearby sources not explicitly modeled.11

10 For a new major stationary source, this includes any direct/precursor pollutant with the potential to emit greater than or equal to the SER and for a modification to an existing major stationary source any direct/precursor pollutant for which the modification results in a significant emissions increase and a significant net emissions increase. 11 The emissions impact of any nearby source that has received a permit but is not yet operational should be included in the air quality assessment. In such cases, consultation with the appropriate permitting authority on the appropriate assessment approach is recommended.

15 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

Once all of these appropriate direct and/or precursor emissions impacts are taken into

account, the estimated cumulative impact is then compared to the NAAQS to determine if there

is a modeled violation. If not, then the NAAQS compliance demonstration is sufficient. If there

are projected NAAQS violations, then the impacts of the emissions increase from the new or

modifying source at those locations are compared to the appropriate SIL to determine whether

that increase will cause or contribute to a violation of the NAAQS. Several aspects of the

cumulative impact analysis for O3 and PM2.5 will be comparable to analyses conducted for other

criteria pollutants, while other aspects will differ due to the issues identified earlier.

II.5.2 PM2.5 PSD Increments Compliance

For PM2.5, where the source impact analysis described in Section II.4 is insufficient to

show that a source will not cause or contribute to a violation of any PM2.5 PSD increment, a cumulative impact analysis is necessary to make the PSD increment demonstration, as described in section 9.2.3 of the 2017 Guideline. A cumulative impact analysis for an increment differs from the NAAQS cumulative impact analysis in that the increment assessment only accounts for the combined impact of the new or modifying source’s emissions increase and certain previous emissions changes from sources (including the modifying source) that affect the PSD increment under the EPA’s PSD regulations. A more complete description of the types of emissions that affect increment consumption and other aspects of the PSD increment system is contained in

Section V.1 of this guidance document. The cumulative impacts are then compared to the appropriate PM2.5 PSD increments to determine whether the new or modifying source emissions

will cause or contribute to a violation of any PM2.5 PSD increment.

16 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

For PM2.5 PSD increments, since the requirement for calculating the amount of

increment consumed was established relatively recently in comparison to the increments for

other pollutants, a new or modified source being evaluated for PM2.5 PSD increments compliance

may still find that it is the first source, or one of only a few sources, with increment-consuming

emissions in a particular attainment or unclassifiable area. As shown in Figure II-2, for such

situations, a permitting authority may have sufficient reason (based on the approach for

conducting source impact analysis described below) to conclude that the impacts of the new or

modified source may be compared directly to the allowable increments, without the need for a

cumulative modeling analysis. This would be the case where it can be shown that any other

increment-consuming sources in the same baseline area, if any, do not have much or any

overlapping impact with the proposed new or modified source.12

Another important consideration for PM2.5 PSD increments is the differences in the EPA

recommended SIL values for Class I and Class II / III areas, as presented in Table II-2. Given

substantially smaller recommended SIL values for Class I areas, there is a greater likelihood that

a proposed new or modifying source would cause or contribute to a PSD increment violation in a

Class I area, even at distances beyond the nominal 50 km near-field application distance.

Section 4.2 of the 2017 Guideline provides screening and compliance assessment approaches for

near-field (50 km or less) and long-range transport (beyond 50 km) situations. The MERPs

Guidance (i.e., Tier 1 Assessment Approach) and the Single-source Modeling Guidance (i.e.,

Tier 2 Assessment Approach) should be referenced for assessing secondary PM2.5 impacts. There is also distance-weighted empirical relationship information (i.e., precursor contributions to

12 The term “increment-consuming source,” as used in this guidance, is intended to refer to any type of source whose emissions changes (increases or decreases) affects the amount of increment consumed or expanded.

17 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

secondary impacts by distance from source) provided within the MERPs Guidance that may be

particularly useful for assessing secondary PM2.5 impacts in long-range transport situations.

Consultation with the appropriate permitting authority and the appropriate EPA Regional Office

is highly recommended for any permit applicants demonstrating long-range Class I area

increment compliance per the requirements of section 4.2.c.ii of the 2017 Guideline.

18 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

III. PSD Compliance Demonstrations for the O3 and PM2.5 NAAQS: Source Impact Analysis

This section provides details regarding the EPA’s recommended approaches for

conducting the source impact analysis as part of a PSD compliance demonstration for the O3

and/or PM2.5 NAAQS.

III.1 O3 NAAQS

This section provides details regarding the EPA’s recommended approaches for conducting the source impact analysis for the O3 NAAQS associated with each of the two

assessment cases presented in Table III-1. In each of the assessment cases, the analysis should

begin by evaluating the impacts of each O3 precursor (VOC and/or NOX) that would be emitted

in a significant amount, i.e., equal to or greater than the respective SER (40 tpy).

Table III-1. EPA Recommended Approaches for Assessing O3 Impacts by Assessment Case

Secondary Impacts Assessment Case Description of Assessment Case Approach*

Case 1: No Air Quality NOX emissions and VOC emissions < 40 tpy SER N/A Analysis

Include each precursor of O3 emitted in a significant amount, see Section II.2. Case 2*: Secondary Air NOX emissions and/or VOC emissions ≥ 40 tpy SER • Tier 1 Approach Quality Impacts (e.g., MERPs) • Tier 2 Approach (e.g., Chemical Transport Modeling)

* In unique situations (e.g., in parts of Alaska where photochemistry is not possible for portions of the year), it may be acceptable for the applicant to rely upon a qualitative approach to assess the secondary impacts. Any qualitative assessments should be justified on a case-by-case basis in consultation with the appropriate permitting authority and the appropriate EPA Regional Office.

19 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

For Case 1, a modeled O3 NAAQS compliance demonstration is not required since

neither O3 precursor (NOX or VOC) is proposed to be emitted in an amount equal to or greater

than the applicable SER. For Case 2, where NOX and/or VOC precursor emissions are greater

than the applicable SER, the permit applicant would need to conduct a compliance demonstration

for secondary O3 impacts for the precursor(s) with emissions equal to or greater than the SER

based on the two-tiered demonstration approach in EPA’s 2017 Guideline.

III.2 PM2.5 NAAQS

This section provides details regarding the EPA’s recommended approaches for

conducting the source impact analysis for the PM2.5 NAAQS associated with each of the four

assessment cases presented in Table III-2. In each of the assessment cases, the analysis should

begin by evaluating the primary PM2.5 impacts of direct emissions that would be emitted in a

significant amount, i.e., equal to or greater than the SER (10 tpy), and each precursor NOX and/or

SO2 that would be emitted in a significant amount, i.e., equal to or greater than the respective

SER (40 tpy).

20 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

Table III-2. EPA Recommended Approaches for Assessing Primary and Secondary PM2.5 Impacts by Assessment Case

Assessment Primary Impacts Secondary Impacts Description of Assessment Case Case Approach Approach*

Include each precursor of PM2.5 emitted in a significant amount, see Case 3*: Appendix W Section II.2. Primary and preferred or Direct PM2.5 emissions ≥ 10 tpy SER Secondary Air approved NOX emissions and/or SO2 emissions ≥ 40 tpy SER • Tier 1 Approach alternative Quality (e.g., MERPs) dispersion model Impacts • Tier 2 Approach (e.g., Chemical Transport Modeling)

Include each precursor of PM2.5 emitted in a significant amount, see Case 4*: Section II.2. Secondary Air Direct PM2.5 emissions < 10 tpy SER N/A Quality NOX emissions and/or SO2 emissions ≥ 40 tpy SER • Tier 1 Approach Impacts Only (e.g., MERPs) • Tier 2 Approach (e.g., Chemical Transport Modeling)

* In unique situations (e.g., in parts of Alaska where photochemistry is not possible for portions of the year), it may be acceptable for the applicant to rely upon a qualitative approach to assess the secondary impacts. Any qualitative assessments should be justified on a case-by-case basis in consultation with the appropriate EPA Regional Office or other applicable permitting authority.

A PM2.5 NAAQS compliance demonstration is not required for Case 1 since neither

direct PM2.5 emissions nor any PM2.5 precursor (NOX or SO2) emissions is proposed to be

emitted in a significant amount. Case 1 is the only assessment case that does not require a

NAAQS compliance demonstration. Each of the remaining three assessment cases would include

21 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

conducting a source impact analysis.

Case 2, where only direct PM2.5 emissions are greater than or equal to the applicable

SER: In this case, the permit applicant may be able to demonstrate that primary PM2.5 impacts

from the proposed increase in direct PM2.5 emissions are below an appropriate SIL based on dispersion modeling using AERMOD or another appropriate preferred model listed in Appendix

A of the 2017 Guideline, or an alternative model subject to the provisions of section 3.2 of the

2017 Guideline.

Case 3, where direct PM2.5 emissions and NOX and/or SO2 precursor emissions are greater than or equal to the applicable SER: In this case, consistent with Case 2, the primary

PM2.5 impacts from direct PM2.5 emissions can be estimated based on application of AERMOD

or an approved alternative model. However, AERMOD does not account for secondary

formation of PM2.5 associated with the source’s precursor emissions. Since the source also

proposes to emit quantities of one or both PM2.5 precursors in significant amounts, an assessment

of their potential impact on secondary PM2.5 is necessary. The assessment of NOX and/or SO2

precursor emission impacts on secondary PM2.5 formation should be conducted based on the

two-tiered demonstration approach in EPA’s 2017 Guideline.

Case 4, where only NOX and/or SO2 precursor emissions are greater than or equal to the

applicable SER: In this case, since direct PM2.5 emissions are insignificant, i.e., below the

applicable SER, the analysis would only address the secondary PM2.5 impacts from NOX and/or

SO2 precursor emissions. Similar to Case 3, the assessment of the precursor emission impacts on

secondary PM2.5 formation for Case 4 would be conducted based on the two-tiered

demonstration approach in EPA’s 2017 Guideline.

22 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

III.3 Assessing Primary PM2.5 Impacts

The assessment of primary PM2.5 impacts from the proposed new or modifying source is

generally the same for the PM2.5 NAAQS and PSD increments. Section 4.2.3.5 of the 2017

Guideline identifies the AERMOD modeling system as the preferred model for addressing direct

PM2.5 emissions unless another preferred model listed in the Guideline is more appropriate, such

as the Offshore and Coastal Dispersion Model (OCD), or the use of an alternative model is

justified consistent with section 3.2 of the 2017 Guideline.

The AERMOD modeling system includes the following regulatory components:

• AERMOD: the dispersion model (U.S. EPA, 2019b);

• AERMAP: the terrain processor for AERMOD (U.S. EPA, 2018b); and

• AERMET: the meteorological data processor for AERMOD (U.S. EPA, 2019c).

Other components that may be used, depending on the application, are:

• BPIPPRIME: the building input processor (U.S. EPA, 2004);

• AERSURFACE: the surface characteristics processor for AERMET (U.S. EPA, 2008);

• AERSCREEN: a screening version of AERMOD (U.S. EPA, 2016b; U.S. EPA, 2011a);

and

• AERMINUTE: a pre-processor to calculate hourly average winds from ASOS 2-minute

observations (U.S. EPA, 2015).

Before applying AERMOD, the applicant should become familiar with the user’s guides associated with the modeling components listed above and the most recent version of the

AERMOD Implementation Guide (U.S. EPA, 2019d). In addition to these documents, detailed guidance on the use of the AERMOD modeling system for estimating primary PM2.5 impacts is

23 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

provided in Appendix B. Because AERMOD is limited to modeling direct PM2.5 emissions,

additional or alternative approaches are used to provide an assessment of secondary PM2.5

impacts from the proposed new or modifying source, as discussed in more detail in the following

sections.

III.4 Assessing O3 and Secondary PM2.5 Impacts

This section provides more detail on the EPA’s recommended approaches for assessing

the impacts of precursor emissions on O3 and/or secondary PM2.5 formation.

III.4.1 Conceptual Model

Each NAAQS compliance demonstration is unique and may require multiple factors to be

considered and assumptions to be thoroughly justified as a part of the technical assessment. A

well-developed modeling protocol that includes a detailed conceptual description of the current

air pollutant concentrations in the area (see Appendix A for examples of elements of a conceptual description) and of the nature of the emissions sources within proximity of the new or modifying emissions source is essential for determining the necessary components of an

13 acceptable assessment of the impact from O3 and/or secondary PM2.5 formation. With timely

13 For more detailed information on the development of such conceptual descriptions for an area, please refer to the following:

Chapter 10 of “Particulate Matter Assessment for Policy Makers: A NARSTO Assessment.” P. McMurry, M. Shepherd, and J. Vickery, eds. Cambridge University Press, Cambridge, England (NARSTO, 2004).

Section 11, “How Do I Get Started? 'A Conceptual Description'” of “Guidance on the Use of Models and Other Analyses for Demonstrating Attainment of Air Quality Goals for Ozone, PM2.5, and Regional Haze.” U.S. Environmental Protection Agency, Research Triangle Park, North Carolina (U.S. EPA, 2007a).

In addition, relevant regional examples include: “Conceptual Model of PM2.5 Episodes in the Midwest,” January 2009, Lake Michigan Air Directors Consortium; and “Conceptual Model of Particulate Matter Pollution in the California San Joaquin Valley,” Document Number CP045-1-98, September 8, 1998.

24 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

and appropriate consultation between the applicant and the appropriate permitting authority, along with the submittal and subsequent approval, if required, of the modeling protocol by the appropriate permitting authority, many potential problems and unintended oversights in the technical assessment can be resolved early in the process or avoided all together.

In the development of an appropriate conceptual description to support an assessment, it is important to fully characterize the current O3 and/or PM2.5 concentrations in the region where

the new or modifying source is to be located and not just the most current design values, which

historically has been used as used as background concentrations in a cumulative modeling

demonstration. For O3, this characterization should take into consideration episodic high O3

concentrations and any trends in the area. For PM2.5, this characterization should take into

consideration the seasonality and speciated composition of the current PM2.5 concentrations and

any long-term trends that may be occurring. It may also be important to describe the typical background concentrations of certain chemical species that participate in the photochemical reactions that form O3 and secondary PM2.5. It is possible that there are mitigating factors for

secondary PM2.5 formation given limitations of other chemical species important in the

photochemical reactions, e.g., minimal NH3 in the ambient environment that could limit any

precursor pollutant from readily reacting to form secondary PM2.5. This understanding of the

atmospheric environment will provide important insights on the potential for secondary

formation and highlight aspects that will need to be accounted for in the source impact and/or

cumulative impact assessment.

A good conceptual description will also characterize the meteorological conditions that

are representative of the region and are associated with periods and/or seasons of higher and

lower ambient O3 and/or 24-hour PM2.5 concentrations. For example, identification of

25 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

meteorological phenomena that typically occur during periods of high daily 8-hour O3 or 24-hour

PM2.5 concentrations, such as low-level temperature inversions, stagnant high pressure systems,

low-level jets, etc., can be extremely important in understanding the importance, or lack thereof,

of photochemistry and secondary PM2.5 formation for the higher ambient O3 and PM2.5

concentrations. The analysis and understanding of meteorological conditions will also inform the

assessment of high O3 episodes and seasonal 24-hour PM2.5 concentrations in the region.

III.4.2 Tier 1 Assessment Approach

As discussed in the section 5.2 of the 2017 Guideline, the EPA has determined that

advances in chemical transport modeling science make it reasonable to provide more specific,

generally-applicable guidance that identifies particular models or analytical techniques that may

be appropriate for use under specific circumstances for assessing the impacts of an individual

proposed source on O3 and secondary PM2.5 concentrations. There is not a preferred model or

technique for estimating O3 or secondary PM2.5 for specific source impacts. Instead, for assessing

secondary pollutant impacts from individual proposed sources, the degree of complexity required

to appropriately assess potential single-source impacts varies depending on the nature of the source, its proposed emissions, and the background environment. In order to provide the user community flexibility in estimating single-source secondary pollutant impacts, which allows for different approaches to credibly address these different areas, the 2017 Guideline recommends a two-tiered demonstration approach for addressing single-source impacts on ambient concentrations of O3 and secondary PM2.5.

26 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

To inform a Tier 1 assessment,14 the existing air quality model-based information that is

used should be appropriate in terms of representing the type of source, its precursor emissions,

and its geographic location, in addition to those elements of the conceptual description discussed

above. The air quality modeling information may be available from past or current SIP

attainment demonstration modeling, published modeling studies, or peer-reviewed literature with estimates of model responsiveness to precursor emissions in contexts that are relevant to the new or modifying source. The estimates of model responsiveness, such as impact on O3

concentrations per ton of NOX or impact on PM2.5 concentrations per ton of SO2 emissions, could then be used in conjunction with the precursor emissions estimates for the proposed new or modifying source to provide a quantitative estimate of the impact of such precursor emissions on the formation of O3 and/or secondary PM2.5 concentrations. The estimates of responsiveness should be technically credible in representing such impacts and it may be advisable for the estimate to reflect an upper bound of potential impacts.

To assist in the development of appropriate Tier 1 demonstration tools, the EPA

developed the MERPs Guidance to provide a framework for permitting authorities to develop

area-specific MERPs. The MERPs Guidance illustrates how permitting authorities may

appropriately develop MERPs for specific areas and use them as a Tier 1 compliance

demonstration tool for O3 and secondary PM2.5 under the PSD permitting program. The MERPs

guidance also addresses the appropriate use of MERPs to reflect the combined ambient impacts

14 A Tier 1 assessment involves the use of technically credible relationships between precursor emissions and a source’s secondary impacts, e.g., as demonstrated in modeling for a source impact analysis, that may be published in the peer-reviewed literature, developed from modeling that was previously conducted for an area by a source, a governmental agency, or some other entity and that is deemed sufficient for evaluating a proposed source’s impacts, or generated by a peer-reviewed reduced form model. In such cases, the EPA expects that existing air quality model- based information regarding the potential for NOX and VOC precursor emissions to form O3 and for SO2 and NOX precursor emissions to form secondary PM2.5 concentrations may be used to establish an appropriate estimate of O3 and/or secondary PM2.5 impacts from the proposed new or modifying source.

27 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

across O3 or PM2.5 precursors and, in the case of PM2.5, the combined primary and secondary

ambient impacts. Such an approach includes flexibility with respect to the use of Tier 1

demonstration tools to generate information relevant for specific regions or areas and

representative of secondary formation in a particular region or area.

Specifically, the MERPs Guidance provides information about how to use CTMs to

estimate single-source impacts on O3 and secondary PM2.5 and how such model simulation

results for specific areas can be used to develop empirical relationships between a source’s O3

and PM2.5 precursor emissions and its secondary impacts that may be appropriate for use as a

Tier 1 demonstration tool. It also provides results from EPA photochemical modeling of a set of

more than 100 hypothetical sources across geographic areas and source types that may be used in

developing MERPs as discussed in the guidance. This flexible and scientifically credible

approach allows for the development of area-specific Tier 1 demonstration tools that better

represent the chemical and physical characteristics and secondary pollutant formation within that

region or area.

As discussed in the MERPs Guidance, the EPA’s Single-source Modeling Guidance provides information to stakeholders about how to appropriately address the variety of chemical and physical characteristics regarding a project scenario and key receptor areas in conducting photochemical modeling to inform development of MERPs. The development of MERPs for O3

and secondary PM2.5 precursors is just one example of a suitable Tier 1 demonstration tool. The

EPA will continue to engage with the modeling community to identify credible alternative

approaches for estimating single-source secondary pollutant impacts, which provide flexibility

and are less resource intensive for PSD permit demonstrations.

As an example, a Tier 1 assessment of secondary O3 and PM2.5 impacts was developed by

28 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

a permit applicant, the Tennessee Valley Authority (TVA), for a major modification at their

Gleason facility in Tennessee in 2018. The TVA and the Tennessee Department of Environment

and Conservation (TDEC) worked closely with EPA Region 4 to ensure that the ambient impacts

analysis was technically sound and consistent with applicable PSD regulations and EPA

guidance. The PSD air quality modeling analysis was submitted to TDEC in late 2018 using an

approach that was consistent with the EPA’s MERPs Guidance to relate facility emissions to

potential downwind impacts of secondary O3 and PM2.5. A more detailed discussion of the

TVA’s technical assessment is provided in Appendix C.

The National Association of Clean Air Agencies (NACAA) Workgroup final report

(NACAA, 2011) provides details on potential approaches to quantify the secondary PM2.5 impacts from a proposed new or modifying source that may be appropriate to inform a Tier 1 assessment of PM2.5 impacts (see Appendix C and D of NACAA, 2011). One suggested method

in the final report is to convert emissions of precursors into equivalent amounts of direct PM2.5

emissions using “pollutant offset ratios” and then use a dispersion model to assess the impacts of

the combination of direct PM2.5 emissions and the equivalent direct PM2.5 emissions. The

“pollutant offset ratios” referenced in that final report were those put forth by the EPA in the

2008 “Implementation of the New Source Review (NSR) Program for Particulate Matter Less

Than 2.5 Micrometers (PM2.5)” final rule (73 Fed. Reg. 28321) concerning the development and

adoption of interpollutant trading (offset) provisions for PM2.5 under state nonattainment area

15 NSR programs for PM2.5. The EPA’s July 23, 2007, technical analysis titled “Details on

15 In the preamble to the 2008 final rule (73 FR 28321), the EPA included preferred or presumptive offset ratios, applicable to specific PM2.5 precursors that state/local air agencies may adopt in conjunction with the new interpollutant offset provisions for PM2.5, and for which the state could rely on the EPA's technical work to demonstrate the adequacy of the ratios for use in any PM2.5 nonattainment area. In a July 21, 2011 memorandum, EPA changed its policy and stated that it no longer supported the ratios provided in the preamble to the 2008 final

29 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

Technical Assessment to Develop Interpollutant Trading Ratios for PM2.5 Offsets,” describes the method used to establish the original "preferred" precursor offset ratios (U.S. EPA, 2007b).

We do not support using the specific results from the EPA's 2007 technical assessment in this context without additional technical demonstration specific to the source(s) and area(s) for which the ratios would be applied. As described in the EPA’s July 21, 2011 memorandum addressing reconsideration of the “preferred” interpollutant offset trading ratios included in the preamble to the 2008 final rule, the EPA acknowledged that existing models and techniques are adequate to “conduct local demonstrations leading to the development of area-specific ratios for

PM2.5 nonattainment areas” and provided a general framework for efforts that may be relevant in

developing appropriate “pollutant offset ratios” for use in hybrid qualitative/quantitative

assessment of secondary PM2.5 impacts (U.S. EPA, 2011b). In the context of PSD compliance

demonstrations, a similar general framework is embodied in the MERPs Guidance in which the

EPA addresses how to conduct modeling to inform the development of a MERP for a particular

area.

The EPA also notes that the NACAA Workgroup “considered, but rejected, other

methods for assessing secondary PM2.5 impacts, including use of a simple emissions divided by

distance (Q/D) metric and use of AERMOD with 100 percent conversion of SO2 and NOX

concentrations to (NH4)2SO4 and (NH4)NO3.” The EPA has reviewed the detailed discussion

provided in Appendix E of the NACAA Workgroup final report and agrees with these

conclusions.

rule as presumptively approvable ratios for adoption in SIPs containing nonattainment NSR programs for PM2.5. Memorandum from Gina McCarthy, Assistant Administrator, to Regional Air Division Directors, “Revised Policy to Address Reconsideration of Interpollutant Trading Provisions for Fine Particles (PM2.5)” (U.S. EPA, 2011b).

30 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

III.4.3 Tier 2 Assessment Approach

As discussed in the 2017 Guideline, a Tier 2 assessment involves application of more sophisticated, case-specific CTMs in consultation with the appropriate permitting authority and conducted consistent with the recommendations in the most current version of the Single-source

Modeling Guidance. Where it is necessary to estimate O3 and/or secondary PM2.5 impacts with case-specific air quality modeling, a candidate model should be selected for estimating single- source impacts on O3 and/or secondarily formed PM2.5 that meets the general criteria for an

“alternative model” where there is no preferred model as outlined in section 3.2.2.e of the 2017

Guideline. The general criteria include:

i. The model has received a scientific peer review;

ii. The model can be demonstrated to be applicable to the problem on a theoretical

basis;

iii. The databases that are necessary to perform the analysis are available and

adequate;

iv. Appropriate performance evaluations of the model have shown that the model is

not biased toward underestimates; and

iv. A protocol on methods and procedures to be followed has been established.

Section 3.2.2 further provides that the appropriate EPA Regional Office, in consultation with the

EPA Model Clearinghouse, is authorized to approve a particular model and approach as an alternative model application.

Both Lagrangian puff models and photochemical grid models may be appropriate for this purpose where those models satisfy alternative model criteria in section 3.2.2 of the 2017

Guideline. That said, the EPA believes photochemical grid models are generally most

31 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

appropriate for addressing O3 and secondary PM2.5 impacts because they provide a spatially and

temporally dynamic realistic chemical and physical environment for plume growth and chemical

transformation. Publicly available and documented Eulerian photochemical grid models such as

the Comprehensive Air Quality Model with Extensions (CAMx) (Ramboll Environ, 2018) and the Community Multiscale Air Quality (CMAQ) (Byun and Schere, 2006) model treat

emissions, chemical transformation, transport, and deposition using time and space variant

meteorology. These modeling systems include primarily emitted species and secondarily formed

pollutants such as O3 and PM2.5 (Chen et al., 2014; Civerolo et al., 2010; Russell, 2008; Tesche

et al., 2006). In addition, these models have been used extensively to support O3 and PM2.5 SIPs and to explore relationships between inputs and air quality impacts in the United States and elsewhere (Cai et al., 2011; Civerolo et al., 2010; Hogrefe et al., 2011).

On August 4, 2017, the EPA released a memorandum (U.S. EPA, 2017b) providing information specific to how the CAMx and the CMAQ model systems were relevant for each of these elements. This memorandum provides an alternative model demonstration for the CAMx and CMAQ photochemical transports models establishing their fit for purpose in PSD compliance demonstrations for O3 and PM2.5 and in NAAQS attainment demonstrations for O3,

PM2.5 and Regional Haze. The memorandum also provides for their general applicability for use

in PSD compliance demonstrations; however, it does not replace the need for such

demonstrations to provide model protocols describing model application choices or the

evaluation of model inputs and baseline predictions against measurements relevant for their

specific use by permit applicants and state, local, and tribal air agencies.

For those situations where a refined Tier 2 demonstration is necessary, the EPA has also

provided the Single-source Modeling Guidance that provides recommended, credible procedures

32 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

to estimate single-source secondary impacts from sources for permit related assessments.

Extensive peer-reviewed literature demonstrates/documents that photochemical grid models have

been applied for single-source impacts and that the models adequately represent secondary

pollutant impacts from a specific facility, in comparison to near-source downwind in-plume measurements. The literature shows that these models can clearly differentiate impacts of a specific facility from those of other sources (Baker and Kelly, 2014; Zhou et al., 2012). Other peer-reviewed research has clearly shown that photochemical grid models are able to simulate impacts from single sources on secondarily-formed pollutants (Baker et al., 2015; Bergin et al.,

2008; Kelly et al., 2015). Further, single-source secondary impacts have been provided in technical reports that further support the utility of these tools for single-source scientific and regulatory assessments (ENVIRON 2012a; ENVIRON 2012b; Yarwood et al., 2011). The EPA firmly believes that the peer-reviewed science clearly demonstrates that photochemical grid models can adequately assess single-source impacts. The EPA recognizes that ongoing evaluations in this area will lead to continual improvements in science and associated predictive capabilities of these models.

For the purposes of conducting a Tier 2 assessment, the application of a CTM will involve case-specific factors that should be part of the consultation process with the appropriate permitting authority and reflected in the agreed-upon modeling protocol. Consistent with the

Single-source Modeling Guidance and section 9.2.1 of the 2017 Guideline, EPA recommends that the modeling protocols for this purpose should include the following elements:

33 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

1. Overview of Modeling/Analysis Project

• Participating organizations

• Schedule for completion of the project

• Description of the conceptual model for the project source/receptor area

• Identify how modeling and other analyses will be archived and documented

• Identify specific deliverables to the appropriate permitting authority

2. Model and Modeling Inputs

• Rationale for the selection of air quality, meteorological, and emissions models

• Modeling domain

• Horizontal and vertical resolution

• Specification of initial and boundary conditions

• Episode selection and rationale for episode selection

• Rationale for and description of meteorological model setup

• Basis for and development of emissions inputs

• Methods used to quality assure emissions, meteorological, and other model inputs

3. Model Performance Evaluation

• Describe ambient database(s)

• Describe evaluation procedures and performance metrics

As stated previously, we expect that the EPA Regional Offices, with assistance from the

OAQPS, may assist reviewing authorities, as necessary, to structure appropriate technical demonstrations leading to the development of appropriate chemical transport modeling applications for the purposes of estimating potential O3 and/or secondary PM2.5 impacts.

34 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

III.5 Comparison to the SIL

This section provides recommendations for source impact analyses where a permit

applicant compares the proposed source’s ambient O3 or PM2.5 impacts to an appropriate SIL as

part of the required demonstration that a proposed source or modification will not cause or

contribute to a violation of the O3 or PM2.5 NAAQS. These recommendations are also generally

applicable for demonstrations that a proposed source or modification will not cause or contribute

to a violation of the PM2.5 PSD increments, see Section V.4. The EPA’s recommended SIL

values for O3 and PM2.5 NAAQS and PM2.5 PSD increments are listed in Table II-1 and Table II-

2. (U.S. EPA 2018a).

III.5.1 SIL Comparison for O3

For Assessment Case 2, an analysis of secondary O3 impacts would be conducted where

the proposed source’s precursor emissions of NOX and/or VOC are equal to or greater than the

respective SERs. The EPA recommends that the assessment of the precursor emission impacts on

O3 formation should be conducted based on the two-tiered demonstration approach as provided

for specific to O3 in section 5.3 of the 2017 Guideline. Under the Tier 1 approach, for source impact analyses, the highest of the multi-season (or episode) averages of the maximum modeled daily 8-hour O3 concentrations predicted each season (or episode) should be compared to the appropriate O3 SIL, since this metric represents the maximum potential daily 8-hour O3 impact

from the proposed source or modification. Under the Tier 2 approach, where a CTM is directly

applied to estimate the source impacts, the comparison should be done at each receptor, i.e., each

modeled grid cell. If the source impact is less than the SIL, then the analysis is generally

sufficient to support a finding that the source will not cause or contribute to a NAAQS violation.

35 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

However, if the source impact is equal to or greater than the SIL, then the analysis is insufficient to show that a source will not cause or contribute to a violation of the NAAQS and a cumulative impact assessment would be necessary.

III.5.2 SIL Comparison for PM2.5

For Assessment Case 2, an analysis of primary PM2.5 impacts would be conducted where the proposed source’s direct PM2.5 emissions are equal to or greater than the applicable SER (10 tpy). In such situations, the modeled estimates of ambient primary PM2.5 concentrations due to direct emissions using the EPA preferred AERMOD dispersion model (or other acceptable preferred or approved alternative model) should be compared to an appropriate PM2.5 SIL in the source impact analysis. The dispersion modeling methods here are similar to the methods used for other primary pollutants, including the use of maximum allowable emissions, following

Table 8-2 of the 2017 Guideline. However, due to the form of the PM2.5 NAAQS, we recommend that one of the following be compared to the SIL, depending on the meteorological data used in the analysis:

• The highest of the 5-year averages of the maximum modeled annual 24-hour PM2.5

concentrations (for the 24-hour PM2.5 NAAQS) or highest of the 5-year averages of

the annual average PM2.5 concentrations (for the annual PM2.5 NAAQS) predicted

each year at each receptor, based on 5 years of representative National Weather

Service (NWS) data;

• The highest modeled 24-hour PM2.5 concentration (for the 24-hour PM2.5 NAAQS) or

the highest modeled average PM2.5 concentration (for the annual PM2.5 NAAQS)

predicted at each receptor based on 1 year of site-specific meteorological data; or the

36 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

highest of the multi-year averages of the maximum modeled annual 24-hour PM2.5

concentration (for the 24-hour PM2.5 NAAQS) or the highest of the multi-year

averages of the maximum modeled annual average PM2.5 concentrations (for the

annual PM2.5 NAAQS) predicted each year at each receptor, based on 2 or more

years, up to 5 complete years, of available site-specific meteorological data; or

• The highest of the 3-year averages of the maximum modeled annual 24-hour PM2.5

concentrations (for the 24-hour PM2.5 NAAQS) or highest of the 3-year averages of

the annual average PM2.5 concentrations (for the annual PM2.5 NAAQS) predicted

each year at each receptor, based on 3 years of prognostic meteorological data.

These metrics represent the maximum potential 24-hour or annual PM2.5 impacts from the proposed source or modification at any receptor, given the form of the NAAQS, and, therefore, provide an appropriate part of the basis for determining whether a cumulative modeling analysis would be needed. If the source impact is less than the SIL, then the analysis is generally sufficient to support a finding that the source will not cause or contribute to a NAAQS violation.

For Assessment Case 3, analyses of both primary and secondary PM2.5 impacts are necessary because the proposed source’s direct PM2.5 emissions and emissions of at least one

PM2.5 precursor are equal to or greater than the respective SERs. In this case, both the primary and secondary PM2.5 impacts from the proposed source or modification would be included in the comparison to the appropriate PM2.5 SIL in the source impact analysis. As with Case 2, the ambient impacts due to direct PM2.5 emissions would be estimated using the EPA preferred

37 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

AERMOD dispersion model (or other acceptable preferred or approved alternative model). For

the assessment of the precursor emission impacts on PM2.5 formation, the EPA recommends that this part of the assessment should be conducted based on the two-tiered demonstration approach as provided for specific to PM2.5 in section 5.4 of the 2017 Guideline. However, the comparison

to the SIL will depend on the type of assessment conducted for the secondary PM2.5 impacts from

the source.

In the SIL comparison for Case 3, the primary and secondary PM2.5 impacts may be

combined in various ways that may entail greater or lesser degrees of conservatism. For example,

combining the peak estimated primary PM2.5 impact with the peak estimated secondary PM2.5 impact, unpaired in time and space, would tend to be a conservative estimate of combined impacts since, as noted above, peak impacts associated with a source’s direct PM2.5 and

precursor emissions are not likely well-correlated in time or space. The conservatism associated with combining peak estimated primary and secondary impacts for comparison to a SIL makes this an appropriate initial approach to combining estimated primary and secondary PM2.5

impacts.

Other approaches for combining primary and secondary PM2.5 impacts for comparison to

a SIL for Case 3 will vary based on the degree of temporal and spatial pairing of estimated

primary and secondary PM2.5 impacts. Full temporal and spatial pairing may not be feasible in

many cases, given that the dispersion modeling and chemical transport modeling may be based

on different data periods. Furthermore, full temporal and spatial pairing of primary and

secondary PM2.5 impacts may not be appropriate in many cases because photochemical grid

modeling represents gridded concentration estimates whereas dispersion modeling produces

estimates at discrete receptor locations and because of the limitations in the skill of both the

38 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

dispersion model and the photochemical grid model to accurately predict impacts on a paired in

time and space basis. As a result, consideration of some degree of temporal pairing of primary

and secondary PM2.5 impacts is most appropriate on a seasonal or monthly basis with

considerations of spatial pairing that reflects the general lack of correlation between primary and

secondary impacts, i.e., primary impacts being higher near the source while secondary impacts

being higher at some distance away from the source.

The permitting authority and the permit applicant should thoroughly discuss the details

regarding combining modeled primary and secondary PM2.5 impacts for Case 3 situations and

should reach agreement during the initial review of the modeling protocol. The permitting

authority should ensure that any approach for combining estimated primary and secondary PM2.5

impacts for comparison to a SIL for Case 3 conforms to the recommendations described above

for Case 2 regarding the form of the modeled estimate. Accordingly, the approach should be

based on the highest of the multi-year averages of the maximum modeled 24-hour or annual

PM2.5 concentrations predicted each year at each receptor, which represents the maximum

potential impact from the proposed source or modification.

For Assessment Case 4, an analysis of secondary PM2.5 impacts would be conducted for

the proposed source’s precursor emissions that are equal to or greater than the respective SERs.

For this source impact analysis, under the Tier 1 approach, the highest of the multi-year averages

of the maximum predicted modeled 24-hour or annual PM2.5 concentrations should be compared

to the appropriate PM2.5 SIL since these metrics represent the maximum potential impact from

the proposed source or modification. Under the Tier 2 approach, where a CTM is directly applied

to estimate the source impacts, the comparison should be done at each receptor, i.e., each modeled grid cell.

39 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

This Page Intentionally Left Blank

40 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

IV. PSD Compliance Demonstrations for the O3 and PM2.5 NAAQS: Cumulative Impact Analysis

Where the source impact analysis described in Section III is insufficient to show that a

source will not cause or contribute to a violation of the O3 or PM2.5 NAAQS, a cumulative

impact assessment will then be necessary to determine whether the source complies with the

NAAQS. A cumulative assessment accounts for the combined impacts of the proposed new or

modifying source’s emissions, emissions from other nearby sources, and representative

background levels of O3 or PM2.5 within the modeling domain. The cumulative impacts are then compared to the O3 or PM2.5 NAAQS to determine whether there is a modeled NAAQS

violation. If not, then the NAAQS compliance demonstration is sufficient. If there are modeled

violations, then the source impact at the location of these violations is compared to the

appropriate SIL to determine if the proposed new or modifying source emissions will cause or

contribute to a violation of the NAAQS. This section provides details on conducting an

appropriate cumulative impact assessment for the O3 and PM2.5 NAAQS.

The cumulative impact assessment should include the following components of O3

impacts, as appropriate, for comparison to the NAAQS:

• Proposed new or modifying source

o Impacts on O3 from each precursor (NOX and/or VOC) that is proposed to be emitted in a significant amount, i.e., equal to or greater than the

respective SER (40 tpy)

• Nearby sources

o Impacts on O3 from precursors (NOX and/or VOC) are typically accounted for through representative monitored background

41 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

• Monitored background level of O3 that accounts for O3 impacts from regional

transport and from nearby sources16

PM2.5

The cumulative impact assessment should include the following components of PM2.5

impacts, as appropriate, for comparison to the NAAQS:

• Proposed new or modifying source

o Primary impacts on PM2.5, i.e., from direct PM2.5 emissions that are proposed to be emitted in a significant amount, i.e., equal to or greater

than the SER (10 tpy)

o Secondary impacts on PM2.5 from each precursor (NOX and/or SO2) that is proposed to be emitted in a significant amount, i.e., equal to or greater

than the respective SER (40 tpy)

• Nearby sources

o Primary impacts on PM2.5

o Impacts on PM2.5 from precursors (NOX and/or SO2) are typically accounted for through representative monitored background

• Monitored background level of PM2.5 that accounts for secondary PM2.5 impacts

from regional transport and from nearby sources, and primary PM2.5 impacts from

background sources not included in the modeled inventory, e.g., minor sources17

16 The emissions impact of any nearby source that has received a permit but is not yet operational should be included in the air quality assessment. In such cases, consultation with the appropriate permitting authority on the appropriate assessment approach is recommended. 17 The emissions impact of any nearby source that has received a permit but is not yet operational should be included in the air quality assessment. In such cases, consultation with the appropriate permitting authority on the appropriate assessment approach is recommended

42 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

As with the source impact analysis, the primary impacts of direct PM2.5 emissions from

the proposed new or modifying source and nearby sources in a cumulative impact analysis

should be estimated based on the AERMOD dispersion model (or other acceptable preferred or

approved alternative model). In addition, EPA recommends that the estimate of secondary PM2.5

impacts from the proposed new or modifying source should be conducted based on the two-

tiered demonstration approach described in section 5.2 of the 2017 Guideline. As noted above,

secondary impacts on PM2.5 from regional transport, precursor emissions from nearby sources,

and primary PM2.5 impacts from background sources not included in the modeled inventory

should be accounted for through representative monitored background concentrations.

IV.1 Modeling Inventory

Section 8 of the 2017 Guideline provides the current required and recommended

approaches for characterizing source emissions and developing the O3 and/or PM2.5 modeling

inventory for purposes of NAAQS compliance modeling in PSD air quality demonstrations.

Section 8.2 and Table 8-2 of the 2017 Guideline address the appropriate emissions limit,

operating level, and operating factor to be modeled, which is the maximum allowable emissions

rate for the proposed new or modifying source in most cases and an allowable emissions rate

adjusted for actual operations for any nearby sources. For applications that require the assessment of secondarily formed O3 or PM2.5 through case-specific chemical transport modeling, information regarding the development of the appropriate modeling inventory can be found in the Single-source Modeling Guidance.

Section 8.3.3 of the 2017 Guideline emphasizes the importance of professional judgment in the identification of nearby and other sources “that are not adequately represented by ambient

43 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

monitoring data” that should be included in the modeled emission inventory and identifies “a

significant concentration gradient in the vicinity of the [proposed] source” as a primary criterion

for this selection. Additionally, the 2017 Guideline suggests that “the number of nearby sources

to be explicitly modeled in the air quality analysis is expected to be few except in unusual

situations” and that “[i]n most cases, the few nearby sources will be located within the first 10 to

20 km from the [proposed] source.” The EPA also provided modeling guidance in March 2011

(U.S. EPA, 2011c) that includes a detailed discussion of the significant concentration gradient

criterion. However, several application-specific factors should be considered when determining

the appropriate inventory of nearby sources to include in the cumulative modeling analysis,

including the potential influence of terrain characteristics on concentration gradients and the

availability and adequacy of ambient monitoring data to account for impacts from nearby sources as well as other background sources.

Consistent with the 2017 revisions to the Guideline, the EPA cautions against the application of very prescriptive procedures for identifying which nearby sources should be included in the modeled emission inventory for NAAQS compliance demonstrations, such as the procedures described in Chapter C, Section IV.C.1 of the draft “New Source Review Workshop

Manual” (U.S. EPA, 1990). Our main concern is that following such procedures in a literal and uncritical manner may, in many cases, increase the likelihood of double-counting modeled and monitored concentrations, resulting in cumulative impact assessments that are overly conservative and would unnecessarily complicate the permitting process. The identification of which sources to include in the modeled emissions inventory should be addressed in the modeling protocol and, as necessary, discussed in advance with the permitting authority.

Since modeling of direct PM2.5 emissions has been limited and infrequent, the availability

44 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

of an adequate direct PM2.5 emission inventory for nearby sources may not exist in all cases.

Recommendations for developing PM2.5 emission inventories for use in PSD applications will be

addressed separately, but existing SIP inventories for PM2.5 or statewide PSD inventories of

sources for refined modeling are expected to provide a useful starting point for this effort.

IV.2 Monitored Background

Section 8.3 of the 2017 Guideline provides recommendations for determination of

monitored background concentrations to include in cumulative impact assessments for NAAQS

compliance, which should account for impacts from existing sources that are not explicitly

included in the modeled inventory and natural sources. From newly-acquired pre-construction

monitoring data and/or existing representative air quality data gathered for purposes of a

permitting analysis, permit applicants should assess and document what the background

monitoring data represent to the extent possible, including any information that may be available

from the state or other agency responsible for siting and maintaining the monitor.18

Determining the monitored background concentrations of O3 and/or PM2.5 to include in

the cumulative impact assessment may entail different considerations from those for other

criteria pollutants lacking secondary formation. An important aspect of the monitored

background concentration for O3 or PM2.5 is that the ambient monitoring data should in most

cases account for the impact of secondary formation of either pollutant from precursor emissions

of existing sources impacting the modeling domain. Additionally, for PM2.5, ambient monitoring

18 Please note in the case of an existing source seeking a permit for a modification, there is potential overlap across secondary impacts from monitored background and from precursor emission from the existing source. In such cases, recommendations for excluding monitored values when the source in question is impacting the monitor in section 8.3.2.b of the 2017 Guideline may need to be modified to avoid overcompensating in cases where the monitored concentrations are also intended to account for the existing source’s impacts on secondary PM2.5.

45 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

data should account for the component of the background levels of primary PM2.5 from

emissions of nearby sources that are not included in the modeled inventory. As with other criteria

pollutants, consideration should also be given to the potential for some double-counting of the impacts from modeled emissions that may be also included in the background monitored concentrations. This should generally be of less importance than the representativeness of the monitor for secondary formation of O3 and PM2.5., unless the monitor is located relatively close to nearby sources of primary PM2.5 that could be impacting the monitor. Also, due to the nature

of O3 and secondary PM2.5, monitored background concentrations of O3 and PM2.5 are more likely to be homogeneous across the modeling domain in most cases compared to most other pollutants.

Depending on the nature of local PM2.5 levels within the modeling domain, it may be

appropriate to account for seasonal variations in monitored background PM2.5 levels, which may

not be correlated with seasonal patterns of the modeled primary PM2.5 levels. For example,

maximum modeled primary PM2.5 impacts associated with low-level emission sources are likely to occur during winter months due to longer periods of stable atmospheric conditions, whereas maximum ambient levels of secondary PM2.5 typically occur during spring and summer months

due to high levels of sulfates (particularly in the eastern United States). The use of temporally-

varying monitored background concentrations in a cumulative impact analysis is discussed in

more detail in Section IV.3.

IV.3 Comparison to the NAAQS

As indicated in Figure II-1, the first step of a cumulative impact analysis consists of a

comparison of the combined modeled and monitored concentrations, as discussed above, with

46 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

the applicable NAAQS to determine if there are any projected violations of the O3 and/or PM2.5

NAAQS.

Ozone differs from other criteria pollutants because it is secondarily formed by NOx and

VOC precursor emissions and there are not direct O3 emissions to be considered in the NAAQS

compliance demonstration. The O3 design value that is representative for the area, rather than the overall maximum monitored background concentration, should generally be used as the monitored component of the cumulative analysis. The O3 design value is based on the 3-year average of the annual fourth-highest daily maximum 8-hour average O3 concentrations (80 FR

65292).

The EPA recommends that the modeled O3 impacts should be added to the monitor-based design value for comparison to the NAAQS, as appropriate. The monitoring data should be representative in that it accounts for O3 formation associated with existing sources both within

and outside of the modeling domain. The EPA recommends that modeled O3 impacts should be

based on a Tier 1 or 2 assessment that accounts for the source’s precursor emissions of NOx

and/or VOC that are proposed to be emitted in a significant amount. The resulting cumulative O3

concentrations would then be compared to the O3 NAAQS (0.070 ppm).

PM2.5

Combining the modeled and monitored concentrations of PM2.5 for comparison to the 24-

hour or annual PM2.5 NAAQS entails considerations that differ from those for other criteria

pollutants due to the issues identified at the end of Section IV.2. Based on assessment cases

shown in Table III-2, the discussion below addresses comparisons to the NAAQS in the context

of dispersion modeling of direct PM2.5 emissions only (i.e., Case 2) and for applications

47 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

involving assessments of secondary PM2.5 impacts (i.e., Cases 3 and 4).

Given the importance of secondary formation of PM2.5 and the potentially high

background levels relative to the PM2.5 NAAQS, greater emphasis is generally placed on the

monitored background levels relative to the modeled inventory for PM2.5 than for other

pollutants. This is true for both PM2.5 NAAQS and PSD increments assessments. Also, given the

probabilistic form of the PM2.5 NAAQS, careful consideration should be given to how the

monitored and modeled concentrations are combined to estimate the cumulative impact levels.

The PM2.5 design value that is representative for the area, rather than the overall

maximum monitored background concentration, should generally be used as the monitored

component of the cumulative analysis. The PM2.5 design value for the annual averaging period is

based on the 3-year average of the annual average PM2.5 concentrations, while the PM2.5 design

value for the 24-hour averaging period is based on the 3-year average of the annual 98th

percentile 24-hour average PM2.5 concentrations (78 FR 3086). Details regarding the

determination of the annual 98th percentile monitored 24-hour value based on the number of days sampled during the year are provided in the data interpretation procedures for the PM2.5 NAAQS

in Appendix N to 40 CFR part 50.

It should be noted here that although the monitored design values for the PM2.5 standards

are defined in terms of 3-year averages, this definition does not preempt or alter the 2017

Guideline’s requirement for use of 5 years of representative NWS meteorological data, at least 1

year of site-specific data, or at least 3 years of prognostic meteorological data for purposes of

19 modeling primary emissions of PM2.5. The 5-year average based on use of representative NWS

meteorological data, the average across one or more (up to 5) complete years of available site-

19 See 40 CFR part 51, Appendix W, section 8.4.2.e.

48 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

specific data, or the average across 3 years of prognostic meteorological data serves as an

unbiased estimate of the 3-year average for purposes of modeling demonstrations of compliance

with the NAAQS. Modeling of “rolling 3-year averages,” using years 1 through 3, years 2

through 4, and years 3 through 5 as recommended in the EPA’s SIP Modeling Guidance, is not

required.

For each case, the EPA recommends that the modeled design concentrations of primary

PM2.5 and/or the modeled secondary PM2.5 impacts should be added to the monitor-based design value for comparison to the NAAQS, as appropriate. The primary PM2.5 modeled design

concentration should be based on:

th • The 5-year average of the modeled annual 98 percentile 24-hour PM2.5

concentrations (for the 24-hour PM2.5 NAAQS) or 5-year average of the modeled

annual average PM2.5 concentration (for the annual PM2.5 NAAQS) predicted each

year at each receptor, based on 5 years of representative NWS data;

th • The modeled 98 percentile 24-hour PM2.5 concentrations (for the 24-hour PM2.5

NAAQS) or modeled average PM2.5 concentration (for the annual PM2.5 NAAQS)

predicted at each receptor based on 1 year of site-specific meteorological data, or

the multi-year average of the modeled annual 98th percentile 24-hour PM2.5

concentrations (for the 24-hour PM2.5 NAAQS) or modeled annual average PM2.5

concentration (for the annual PM2.5 NAAQS) predicted each year at each receptor,

based on 2 or more years, up to 5 complete years, of available site-specific

meteorological data; or

th • The 3-year average of the modeled annual 98 percentile 24-hour PM2.5

concentrations (for the 24-hour PM2.5 NAAQS) or 3-year average of the modeled

49 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

annual average PM2.5 concentration (for the annual PM2.5 NAAQS) predicted each

year at each receptor, based on 3 years of prognostic meteorological data.

The EPA recommends that secondary PM2.5 modeled impacts should be based on either a

Tier 1 or 2 assessment accounting for the source’s PM2.5 precursor emissions of NOx and/or SO2

that are proposed to be emitted in a significant amount. The resulting cumulative PM2.5

3 concentrations would then be compared to the 24-hour PM2.5 NAAQS (35 μg/m ) and/or the

3 annual PM2.5 NAAQS (12 μg/m ).

Specifically, for Case 2, where the source’s direct PM2.5 emissions are equal to or greater

than the SER, the modeled design concentration should be based on AERMOD (or other

acceptable preferred or approved alternative model) estimates of the proposed source’s and other

nearby sources’ direct PM2.5 emissions combined with the monitor-based design value. The

monitor should be representative in that it accounts for secondary PM2.5 formation associated

with existing sources both within and outside of the modeling domain, in addition to the

background levels of primary PM2.5 associated with nearby and background sources that are not

included in the modeled inventory.

For Case 3, where the source’s direct PM2.5 emissions and NOX and/or SO2 precursor

emissions are proposed to be emitted in amounts equal to or greater than the respective SERs, the

cumulative impact for comparison to the NAAQS should be based on the sum of the modeled

design concentration for primary PM2.5 impacts (from dispersion model estimates based on the

proposed source’s and other nearby source’s direct PM2.5 emissions), the modeled secondary

PM2.5 impacts (based on a Tier 1 or 2 assessment accounting for the proposed source’s PM2.5 precursor emissions), and the monitored design value (see Case 2 discussion above on monitor representativeness).

50 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

For Case 4, where the source’s NOX and/or SO2 precursor emissions are proposed to be

emitted in amounts equal to or greater than the respective SERs, the cumulative impact for

comparison to the NAAQS should be based on the sum of the modeled secondary PM2.5 impacts

(based on a Tier 1 or 2 assessment accounting for the proposed source’s PM2.5 precursor

emissions) and the monitor-based design value (see Case 2 discussion above on monitor

representativeness).

The recommendations provided above constitute a First Level analysis for PM2.5 NAAQS

compliance demonstrations. For applications where impacts from primary PM2.5 emissions are

not temporally correlated with background PM2.5 levels, combining the modeled and monitored

levels as described above may be overly conservative in some situations. For example, there are

areas of the country where background PM2.5 levels are substantially higher on average during

the summer months as compared to the winter months; however, the projected modeled impacts

from the new or modified source may be substantially greater in the winter rather than in the

summer. In such cases, a Second Level modeling analysis may be advisable to account for these

temporal relationships. Such an analysis would involve combining the monitored and modeled

PM2.5 concentrations on a seasonal (or quarterly) basis, as appropriate. The use of a seasonally-

varying monitored background component is likely to be a more important factor for the 24-hour

PM2.5 NAAQS analysis than for the annual PM2.5 NAAQS. Careful evaluation of when model

projections of PM2.5 impacts and background PM2.5 levels peak throughout the year is

recommended before embarking on a Second Level modeling analysis. This is because the First

Level approach may already adequately capture the temporal correlation. As a part of this

process to determine the appropriate level of analysis, the permit applicant should consult with the appropriate permitting authority and then reflect the appropriate approach in their modeling

51 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

protocol.

The AERMOD model provides several options for specifying the monitored background

concentration for inclusion in the cumulative impact assessment. The options that are most

relevant to PM2.5 analyses include:

• For First Level 24-hour or annual PM2.5 NAAQS analyses, an option to specify a

single annual background concentration that is applied to each hour of the year,

and

• For Second Level 24-hour PM2.5 NAAQS analyses, an option to specify four

seasonal background values that are combined with modeled concentrations on a

seasonal basis.

The AERMOD model also allows the user to track the effect of background concentrations on

the cumulative modeled design concentration.

For Second Level 24-hour PM2.5 NAAQS modeling analyses, EPA recommends that the

distribution of monitored data equal to and less than the annual 98th percentile be appropriately

divided into seasons (or quarters) for each of the three years that are used to develop the

monitored design value. This will result in data for each year of the multi-year data, which contains one season (or quarter) with the 98th percentile value and three seasons (quarters) with

maximum values which are less than or equal to the 98th percentile value. The maximum

concentration from each of the seasonal (or quarterly) subsets should then be averaged across

these three years of monitoring data. The resulting average of seasonal (or quarterly) maximums

should then be included as the four seasonal background values within the AERMOD model.

Therefore, the monitored concentrations greater than the 98th percentile in each of the three years would not be included in the seasonal (or quarterly) subsets. These excluded monitored

52 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

concentrations are the same values that are excluded when determining the monitored design

value. An example of the calculations for a Second Level 24-hour PM2.5 NAAQS modeling

analysis is provided in Appendix D.

For a monitor with a daily (1-in-1 day monitor) sampling frequency and 100% data

completeness, the highest seven monitored concentrations for each year would be excluded from

the seasonal (or quarterly) subdivided datasets. Similarly, for a monitor with every third day (1-

in-3 day monitor) sampling frequency and 100% data completeness, the highest two monitored

concentrations for each year would be excluded from the seasonal (or quarterly) subdivided datasets. The monitored concentrations excluded from the subdivided datasets could primarily come from one or two seasons (or quarters) each year or could be evenly distributed across all four seasons (or quarters) each year. Additionally, the monitored concentrations not included in the subdivided datasets could shift seasonally (or quarterly) from one year to the next. Given the reason for considering a Second Level 24-hour analysis (i.e., lack of temporal correlation between modeled and monitored concentrations), it is likely that the monitored data greater than the 98th percentile would be concentrated in one or two seasons as opposed to evenly distributed

throughout the year. As mentioned earlier, see Appendix N of 40 CFR part 50 in determining the

appropriate 98th percentile rank of the monitored data based on the monitor sampling frequency

and valid number of days sampled during each year.

The EPA does not recommend a "paired sums" approach on an hour-by-hour basis because of the spatial and temporal variability throughout a typical modeling domain on an hourly basis and the complexities and limitations of hourly observations from the current PM2.5

ambient monitoring network. The implicit assumption underlying this “paired sums’ approach is

that the background monitored levels for each hour are spatially uniform and that the monitored

53 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

values are fully representative of background levels at each receptor for each hour. Such an

assumption does not account for the many factors that contribute to the temporal and spatial

variability of ambient PM2.5 concentrations across a typical modeling domain on an hourly

basis.20 Furthermore, the pairing of daily monitored background and 24-hour average modeled concentrations is not recommended except in rare cases of relatively isolated sources where the available 1-in-1 day monitor can be shown to be representative of the ambient concentration levels in the areas of maximum impact from the proposed new source. In most cases, the seasonal (or quarterly) pairing of monitored and modeled concentrations previously described in the Second Level approach should sufficiently address situations in which the impacts from primary PM2.5 emissions are not temporally correlated with background PM2.5 levels. Any

monitor-model pairing approach aside from the First or Second Level methods should be

justified on a case-by-case basis in consultation with the appropriate permitting authority and the

appropriate EPA Regional Office.

IV.4 Determining Whether Proposed Source Causes or Contributes to Modeled Violations

If the cumulative impact assessment following these recommendations results in

predicted modeled violations of the O3 and/or PM2.5 NAAQS, then the permit applicant will need

20 The complexity of the PM2.5 ambient monitoring network presents special challenges with a "paired sum" approach that are not present with other NAAQS pollutants. The Federal Reference Method (FRM) PM2.5 monitoring network is based on 24-hour samples that are taken on average every third day at the 1-in-3 day monitors. The frequency of daily or 1-in-1 day PM2.5 monitors is steadily increasing but is relatively limited to the largest cities and metropolitan regions of the U.S. Various methods to "data fill" the 1-in-3 day monitoring database to create a pseudo-daily dataset have been explored in a few situations, but none of these data filling methods have been demonstrated to create a representative daily PM2.5 dataset that the EPA would consider acceptable for inclusion in a PM2.5 NAAQS compliance demonstration. The use of continuous PM2.5 monitors, which are more limited in number compared to the FRM monitors and may require careful quality assurance of individual hourly measurements, may be an option but should be discussed in advance with the appropriate permitting authority.

54 Does not represent final Agency action; Draft for public review and comment; 02/10/2020 to demonstrate that the proposed source’s emissions do not cause or contribute to the modeled

NAAQS violations. In the SILs Guidance, the EPA explained that the permitting authority may further evaluate whether the proposed source or modification will cause or contribute to predicted violations by comparing the proposed source’s modeled impacts, paired in time and space with the predicted violations, to an appropriate SIL. The proposed source or modification would not be considered to cause or contribute to predicted violations of the O3 or PM2.5

NAAQS where the modeled impacts of the proposed source or modification at those particular times and locations are less than the appropriate O3 or PM2.5 NAAQS SIL. As explained in the

SILs Guidance, a permitting authority that chooses to use an O3 or PM2.5 SIL value to support a

PSD permitting decision should justify the value and its use in the administrative record for the permitting action.

A demonstration that a proposed source or modification does not cause or contribute to a predicted violation should be based on a comparison of the modeled concentrations (primary and secondary impacts) at the receptor location(s) showing the violation(s) of the O3 or PM2.5

NAAQS to the appropriate O3 or PM2.5 NAAQS SIL, i.e.,

• For a predicted violation of the O3 NAAQS, the average of the predicted annual

th (or episodic) 98 percentile daily maximum 8-hour averaged O3 concentrations at

the affected receptor(s) should be compared to an appropriate O3 NAAQS SIL,

e.g., SIL values recommended by EPA in the SILs Guidance (Table II-1).

• For a predicted violation of the annual PM2.5 NAAQS, the average of the

predicted annual concentrations at the affected receptor(s) should be compared to

an appropriate PM2.5 annual NAAQS SIL, e.g., SIL values recommended by EPA

in the SILs Guidance (Table II.1).

55 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

• For a predicted violation of the 24-hour PM2.5 NAAQS, the average of the

predicted annual 98th percentile 24-hour average concentrations at the affected

receptor(s) should be compared to an appropriate PM2.5 24-hour NAAQS SIL,

e.g., SIL values recommended by EPA in the SILs Guidance (Table II-1).

56 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

V. PSD Compliance Demonstration for the PM2.5 Increments

As summarized in Section II of this guidance, CAA section 165(a)(3) requires that

proposed new and modified major stationary sources seeking a PSD permit must demonstrate

that their proposed emissions increases will not cause or contribute to a violation of any NAAQS

or PSD increments. Based on the flow diagram presented in Figure II-2, this section describes the EPA’s recommendations for completing the required compliance demonstration for the PSD increments for PM2.5.

V.1 Overview of the PSD Increment System

This section provides an overview of the PSD increment system by defining basic terms,

such as increment, baseline concentration, baseline area, trigger date, minor source baseline date,

and major source baseline date. This section also introduces and discusses the concepts of

increment consumption and expansion.

V.1.1 PSD Increments and Baseline Concentration

The term “increment” generally refers to what the CAA calls the “maximum allowable

increase over baseline concentrations” with respect to a criteria pollutant. The CAA section

169(4) defines “baseline concentration,” generally, as “the ambient concentration levels which

exist at the time of the first application for a [PSD] permit for an area subject to this part….”21

Accordingly, an increment analysis is generally concerned with the emissions increases affecting

21 EPA’s regulations at 40 CFR 52.21(b)(14)(ii) (and 51.166(b)(14)(ii)) provide that the triggering application is to be a complete PSD application. Hence, the term “complete application” will be used throughout this section with regard to the minor source baseline date and increment consumption.

57 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

air quality in a particular PSD area after the date that the first complete PSD application is submitted to the permitting authority.22 When comparing the ambient impact of such total emissions increases against the increment value for a particular pollutant, a cumulative increase

in the ambient concentration of that pollutant that is greater than the increment generally is

considered “significant deterioration.” When the cumulative impact analysis identifies significant

deterioration in this way, the permitting authority should determine whether the emissions

increase from the proposed new source or modification will cause or contribute to the projected

violation of the PSD increment.

Based on the statutory definition of baseline concentration, as described above, it is

conceptually possible to measure whether there will be significant deterioration in at least two

separate ways. The first way involves comparing a direct modeled projection of the change in air

quality caused by all increment-consuming and expanding emissions to the increment in the area of concern (known as the baseline area, discussed below in Section V.1.2). The second approach is to make a determination of whether the current monitored ambient air quality concentration in the applicable baseline area, supplemented by the modeled impact of the proposed source, will exceed an allowable ambient air quality ceiling. This latter approach requires comparing such monitored concentration(s) to the sum of the increment and the baseline concentration for the baseline area.

Historically, because of the lack of monitoring data to adequately represent the baseline concentration combined with various other limitations associated with the use of ambient air

22 The EPA also considers emissions decreases occurring after the date of the first PSD application to affect increment consumption to the extent that such decreases cause an improvement of air quality in the area of concern. Thus, the concept of increment “expansion” is also discussed in this section.

58 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

quality monitoring data for measuring increment consumption,23 the EPA has recommended that

the required increment analysis be based exclusively on the first approach, which models the

increment-related emissions increases or decreases to determine the resulting ambient air quality change and compares this value with the increments for a particular pollutant.

V.1.2 PSD Baseline Area and Key Baseline Dates

In order to determine whether a PSD increment would be violated as part of a PSD permit

review, it is necessary to identify (1) the affected geographic area in which the increment will be

tracked and (2) the key baseline dates after which emissions changes affect increment in that

area. The relevant geographic area for determining the amount of increment consumed is known

as the “baseline area.” The baseline area is established primarily on the basis of the location of

the first major source to submit a complete PSD application after an established “trigger date”

(see discussion of key dates below) and may be comprised of one or more areas that are

designated as “attainment” or “unclassifiable” pursuant to CAA section 107(d) for a particular

pollutant within a state. In accordance with the regulatory definition of baseline area at 40 CFR

52.21(b)(15), the area is an “intrastate area” and does not include any area in another state.24 At a

minimum, the baseline area is the attainment or unclassifiable area in which the first PSD

applicant after the trigger date proposes to locate, but additional attainment or unclassifiable

areas could be included in a particular baseline area when the proposed source’s modeled impact

23 The EPA described certain limitations associated with the use of ambient air quality monitoring data for measuring increment consumption in the preamble to its proposed PSD regulations in 1979. For example, the CAA provides that certain emissions changes should not be considered increment consuming. These limitations generally continue to apply to the extent that certain emissions changes detected by an ambient monitor are not considered to consume increment. See 44 Fed. Reg. 51924, 51944 (September 5, 1979). 24 While baseline dates are established on an intrastate basis, once a baseline area is established, emissions changes from other states may contribute to the amount of increment consumed.

59 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

in any such additional areas exceeds certain concentrations specified in the regulatory definition

of baseline area. Once a baseline area has been established, subsequent PSD applicants

proposing to locate, or which could have a significant impact, in that area should rely on the

associated baseline dates, discussed below, to determine whether the new or modified source’s

proposed emissions would cause or contribute to an increment violation.

Within any baseline area, three key dates will apply in order to conduct the required

increment analysis: (1) trigger date; (2) minor source baseline date; and (3) major source baseline

date. The trigger date is a date fixed by regulation for each pollutant at 40 C.F.R.

52.21(b)(14)(ii), which is the earliest date after which proposed new or modified major sources

submitting a complete PSD application establishes the “minor source baseline date” in a newly

established baseline area. Accordingly, the minor source baseline date is the date on which PSD

permit applicants must actually begin tracking increment tracking. Depending upon the number

of separate attainment and unclassifiable areas that exist for a particular pollutant in a state and

the timing of major source construction within the state, there may be a number of minor source

baseline dates that apply to different baseline areas established in that state. Beginning with the

PSD source whose complete application has established the minor source baseline date in a

particular area, any increase or decrease in actual emissions occurring after the minor source

baseline date at any source that will affect air quality in the baseline area will affect the amount

of PSD increment consumed in that baseline area (in the case of an emissions decrease, see discussion on increment expansion in Section V.1.3 of this guidance, below).

Finally, the “major source baseline date” is a date fixed by regulation for each pollutant at

52.21(b)(14)(i) and precedes the trigger date. As further explained below, changes in emissions resulting from construction at major stationary sources only that occur after the major source

60 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

baseline date but before the minor source baseline date will also affect increment. The

relationship of these three key dates with each other is further illustrated in Figure V-1.

Figure V-1. Determining Baseline Date(s) and When Increment Consumption Starts

Start Major Source Baseline Date Trigger Date Minor Source Baseline Date

Date when actual emissions associated Earliest date after which the minor source Date when actual emissions changes from with construction at a major source affect baseline date may be established all sources affect the available increment increment

SO2 and PM10 - 01/06/1975 SO2 and PM10 - 08/07/1977 Date of first complete PSD

NOX - 02/08/1988 NOX - 02/08/1988 permit application

PM2.5 - 10/20/2011 PM2.5 - 10/20/2011

Emissions changes occurring before the minor source baseline date generally do not

affect increment in an area (i.e., are not increment-consuming) but are considered to affect the

baseline concentration, which, as explained above, represents the ambient pollutant

concentration levels that exist at the time of the minor source baseline date, or the date of the

first complete application for a PSD permit in a an area after the trigger date. However, as noted

above, the CAA provides an exception for certain emissions changes that occur specifically at

major stationary sources as a result of construction25 that commences after the major source

baseline date. Specifically, for projects at major stationary sources on which construction

commenced on a date prior to the major source baseline date, the changes in emissions from such

projects affect the baseline concentration (not the amount of increment consumed) even if the

emissions change may not actually occur until after the major or minor source baseline dates.

Alternately, for projects at major stationary sources on which construction commenced after the

25 The CAA section 169(2)(C) indicates that the term “construction,” when used in connection with any source or facility, includes modifications defined in CAA section 111(a)(4). “Modification” is defined at section 111(a)(4) to mean any physical change or change in the method of operation at a stationary source which increases the amount of any air pollutant emitted by the source or which results in the emission of any air pollutant not previously emitted.

61 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

major source baseline date, the project emissions will be considered to affect increment, even if

the new or modified source actually begins operation before the minor source baseline date.

V.1.3 PSD Increment Expansion

The “increment consumption” analysis allows permit applicants and permitting authorities to take into account emissions reductions that occur in the baseline area of concern.

Such emissions reductions are generally said to result in the expansion of increment in the area; however, not all emissions reductions truly result in an expansion of the increment. Some emissions reductions, instead, result in a freeing up of increment that had previously been consumed.

In the case of true “increment expansion,” emissions in the area are allowed to increase by the amount allowed by the original increment plus the amount of air quality improvement

(relative to the baseline concentration) achieved by the reduction of emissions that were not

considered to consume increment because of their relationship to the established baseline dates

for the area.26 In such cases, it is appropriate to model the emissions decrease as a negative

amount to account for the resulting lowering of the baseline concentration and simulate the

expansion of the increment.

On the other hand, in cases where a source’s emissions contribute to the amount of

increment consumed, a reduction in such increment-consuming emissions at some later date

26 The concept of increment expansion is derived from CAA section 163(a), which provides that a PSD applicant must assure “that maximum allowable increases over baseline concentrations … shall not be exceeded.” [Emphasis added.] The target for determining significant deterioration thus becomes the ambient concentration resulting from the sum of the increment and the baseline concentration. When a decrease in emissions that contribute to the baseline concentration occurs, an emissions increase that simply “restores” the air quality to the original baseline concentration in a particular baseline area can be allowed, regardless of the amount of increment otherwise being consumed.

62 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

results in some amount of the consumed increment being freed up. That is, the resulting air

quality improvement is now available for a source to increase its emissions within the limits of

the original increment level. A subsequent reduction in increment-consuming emissions should

not be modeled as a negative value to determine the amount of increment that has been freed up; instead, such emissions reductions are simply no longer counted in the increment consumption

equation.

V.2 PSD PM2.5 Increments

In 2010, the EPA established the PM2.5 increments at the levels shown in Table V-1 through the “Prevention of Significant Deterioration (PSD) for Particulate Matter Less Than 2.5

Micrometers (PM2.5) – Increments, Significant Impact Levels (SILs) and Significant Monitoring

Concentration (SMC)” final rule.27 This 2010 rule established October 20, 2011, as the trigger

date and October 20, 2010, as the major source baseline date for PM2.5 increments. The EPA

developed the increment system for PM2.5 generally following the same concepts that were

previously applied for development of the increments for PM10, SO2, and nitrogen dioxide

(NO2). As explained above, the framework reflects the statutory concepts set forth in the statutory definition of baseline concentration that was explained in Section V.1 of this guidance.

Table V-1. PM2.5 Increments Class I Class II Class III Increments, µg/m3 Annual arithmetic mean………………………….……...…..……….………… 1 4 8 24-hour maximum………………………………..…..…………………………. 2 9 18

Source: Prevention of Significant Deterioration (PSD) for Particulate Matter Less Than 2.5 Micrometers (PM2.5) - Increments, Significant Impact Levels (SILs) and Significant Monitoring Concentration (SMC) final rule (75 FR 64864)

27 See 75 FR 64864.

63 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

The obvious difference between an increment analysis and the NAAQS analysis for

PM2.5 is that the increment analysis is concerned with the degree of change in air quality caused

by a new or modified PSD source rather than the impact of that source on overall air quality (as

defined by the applicable NAAQS) in the area of concern (baseline area). With this in mind, it

should be noted here that an increment analysis is relevant only to the extent that NAAQS

compliance has been ensured. That is, an adequate air quality analysis demonstrating compliance

with the statutory requirements must ensure that the proposed PSD source’s emissions will not

cause or contribute to either the NAAQS or PSD increments.28

Another key difference involves the modeling inventory from which the necessary

emissions data is derived. That is, only sources that have PM2.5 emissions (direct and precursor)

that affect the amount of increment consumed in the area of concern should be included in the

modeling inventory for the increment analysis. Moreover, from such sources only those specific emissions changes that affect increment should be included in the actual modeling analysis.

The cumulative impact analysis for PM2.5 increments is also different and based on the

actual emission changes occurring at existing sources in the baseline area after the pertinent baseline dates (i.e., major and minor source baseline dates), whereas NAAQS analyses are generally based on the cumulative impact associated with the maximum allowable emissions from the new or modifying source and other nearby sources (with specific provisions for operating levels of nearby sources). Furthermore, ambient monitoring data, while useful for establishing background concentration for the NAAQS analysis, may not be particularly useful for the typical increment analysis. The limitations associated with using monitoring data for an

28 The CAA section 163(b)(4) provides that the maximum allowable concentration of any air pollutant allowed in an area shall not exceed the concentration allowed by the primary or secondary NAAQS.

64 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

increment analysis are discussed in greater detail in Sections V.1 and V.3 of this guidance.

It is also important to note that the PM2.5 NAAQS and increments for the 24-hour

averaging period are defined in different forms and therefore must be analyzed differently.29 The

th 24-hour PM2.5 NAAQS is defined based on the 3-year average of the annual 98 percentile of the

24-hour average concentrations, while the 24-hour PM2.5 increments are based on the second

highest maximum 24-hour concentration.

V.3 PSD Compliance Demonstration for the PM2.5 Increments

The initial steps for the PM2.5 increment analysis, which include the determination of the

significant emissions increases to include in the source impact analysis and comparison of the

modeled impacts against the PM2.5 SILs will rely upon the results derived from the PM2.5

NAAQS analysis described in Sections III and IV of this guidance. Moreover, the technical

approach involving the options and alternatives agreed upon for estimating secondary PM2.5

impacts and combining primary and secondary PM2;5 impacts for the NAAQS analysis will also

be relevant for completing the PM2.5 increment analysis to determine whether the emissions

increase from the proposed source or modification will cause or contribute to any PM2.5

increment violation.

V.3.1 PM2.5 Increments: Source Impact Analysis

The EPA’s recommendations on completing the required compliance demonstration for the PM2.5 PSD increments is based upon the same four assessment cases detailed in Section II.4

for PM2.5 NAAQS. As shown in Table V-2, a modeled compliance demonstration is not required

29 The annual NAAQS and increments for PM2.5 are both measured as annual arithmetic mean values.

65 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

for Case 1 since neither direct PM2.5 emissions nor PM2.5 precursor (NOX and/or SO2) emissions are equal to or greater than the respective SERs. Case 1 is the only assessment case that does not require a modeled compliance demonstration for PM2.5, whereas each of the remaining three assessment cases would necessitate a source impact analysis that should be conducted following the detailed recommendations provided in previous sections for a NAAQS analysis.

66 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

Table V-2. EPA Recommended Approaches for Assessing Primary and Secondary PM2.5 Impacts by Assessment Case

Assessment Primary Impacts Secondary Impacts Description of Assessment Case Case Approach Approach*

Case 1: Direct PM2.5 emissions < 10 tpy SER No Air Quality N/A N/A NOX emissions and SO2 emissions < 40 tpy SER Analysis Case 2: Appendix W preferred or Primary Air Direct PM2.5 emissions ≥ 10 tpy SER approved N/A Quality NOX emissions and SO2 emissions < 40 tpy SER alternative Impacts Only dispersion model Include each precursor of PM2.5 emitted in a significant amount, see Case 3: Appendix W Section II.2. Primary and preferred or Direct PM2.5 emissions ≥ 10 tpy SER Secondary Air approved NOX emissions and/or SO2 emissions ≥ 40 tpy SER • Tier 1 Approach alternative Quality (e.g., MERPs) dispersion model Impacts • Tier 2 Approach (e.g., Chemical Transport Modeling)

Include each precursor of PM2.5 emitted in a significant amounts, see Case 4: Section II.2. Secondary Air Direct PM2.5 emissions < 10 tpy SER N/A Quality NOX emissions and/or SO2 emissions ≥ 40 tpy SER • Tier 1 Approach Impacts Only (e.g., MERPs) • Tier 2 Approach (e.g., Chemical Transport Modeling) * In unique situations (e.g., in parts of Alaska where photochemistry is not possible for portions of the year), it may be acceptable for the applicant to rely upon a qualitative approach to assess the secondary impacts. Any qualitative assessments should be justified on a case-by-case basis in consultation with the appropriate EPA Regional Office or other applicable permitting authority.

A modeling analysis based solely on the PSD applicant’s proposed emissions increase

(i.e., source impact analysis) that does not predict anywhere an ambient impact equal to or

greater than the applicable PM2.5 SIL generally will satisfy the requirement for a demonstration

that the source will not cause or contribute to a violation of the PM2.5 increments. When the PSD

67 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

applicant relies on such analysis to make the required compliance demonstration, the EPA

recommends that the applicant should include: (1) a comparison of the predicted impacts of the

proposed new or modified source and the allowable increment values, (2) information on the

extent, if any, to which increment has already been consumed since the major source baseline

date (by major source construction occurring prior to the minor source baseline date) or since the

minor source baseline date by nearby emissions changes occurring prior to the proposed source,

and (3) information on increment consumption or expansion by more distant emissions changes.

In light of the relatively recent establishment of the fixed dates (i.e., major source

baseline date and trigger date) associated with the PM2.5 increments (compared to comparable fixed dates for other PSD increments), and the possibility that the minor source baseline date for a particular area has not yet been set, a proposed new or modified source being evaluated for compliance with the PM2.5 increments in a particular area may be the first source in the area with

increment-consuming emissions. As indicated in Figure II-2, under this situation, a permitting

authority may have a sufficient basis to conclude that the PM2.5 impacts of the new or modified

PSD source, although greater than the applicable PM2.5 SILs, may be compared directly to the

allowable PM2.5 increments without the need for a cumulative analysis (described in Section

V.3.2 of this guidance below). Reliance on this initial source impact analysis (rather than a source or cumulative impact analysis that is compared to the applicable PM2.5 SILs) likely would be appropriate to assess the amount of increment consumed when the proposed new or modified source represents the first complete PSD application since the trigger date, thus establishing the

baseline concentration in the area, and there has been no other major source construction since

the major source baseline date.

68 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

V.3.2 PM2.5 Increments: Cumulative Analysis

Where the source impact analysis described above is insufficient to show that a proposed

PSD source will not cause or contribute to a violation of the PM2.5 PSD increments, a cumulative

impact assessment would be necessary to complete the required increment analysis. A

cumulative assessment of increment consumption accounts for the combined impacts of the

following:

1. Direct and/or precursor allowable emissions that the proposed new or modifying

source would emit in significant amounts;

2. Direct and/or precursor actual emissions changes that have occurred at existing

sources (including the existing source at which a major modification is being

proposed, where applicable) since the minor source baseline date for the proposed

source’s baseline area;

3. Direct and/or precursor actual emissions from any major stationary source on which

construction commenced after October 20, 2010 (major source baseline date for

PM2.5); and

4. Direct and/or precursor allowable emissions of permitted sources that are not yet fully

operative.30

Unlike the guidance provided for the cumulative NAAQS analysis for PM2.5, it is not

typically practical to utilize ambient monitoring data to represent any portion of the impacts that

affect the PM2.5 increments. Therefore, it is usually necessary to model the applicable emissions from any existing source that will be considered to consume a portion of the PM2.5 increments in

30 Regarding the use of allowable emissions, see 40 CFR 52.21(b)(21)(iv).

69 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

the baseline area of concern. It is highly recommended that the PSD applicant work closely with

the permitting authority to determine the existing sources (including newly permitted sources) of

direct PM2.5 and precursor emissions that should be included in the modeling inventory for the increment analysis. Sources whose emissions have not changed substantially since the applicable baseline date may not need to be included for purposes of increment consumption. If there is reason to believe that an existing source’s actual emissions have decreased since the applicable baseline date, the PSD applicant may want to check with the permitting authority to ascertain whether the authority allows for increment expansion to be considered.

Once the modeling inventory for the increment analysis has been developed and

approved, and the increment-consuming emissions have been determined, the modeled

cumulative impacts resulting from the increases and decreases in emissions are then compared to

the PM2.5 increments to determine whether any increment violations will result. This section

provides recommendations on conducting an appropriate cumulative impact assessment for

PM2.5 increments.

V.3.2.1 Assessing Primary PM2.5 Impacts

As explained in Section III.3 of this guidance, the assessment of primary PM2.5 impacts

from the proposed new or modifying PSD source is essentially the same for the PM2.5 NAAQS

and increments. In both cases, the permit applicant must account for the impacts from the

proposed new or modifying source’s allowable emissions increase of direct PM2.5.

To assess the impact of direct PM2.5 emissions from existing increment-consuming sources, actual emissions increases that have occurred since the applicable minor source baseline date should generally be modeled. Alternatively, existing source impacts from direct PM2.5

70 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

emissions may be conservatively modeled using an existing source’s allowable emissions where

the PSD applicant determines that such emissions are more readily available and especially when

such allowable emissions are not expected to contribute substantially to the amount of increment

consumed. In the event that an applicant chooses to conduct the cumulative analysis using

allowable emissions and identifies potential problems concerning increment consumption, the

PSD applicant may then rely on more refined data that better represent a particular source’s actual emissions.

The PM2.5 increments analysis would follow the traditional approach involving modeling

only direct PM2.5 emissions changes that affect the increment and should be based on application

of AERMOD (or other acceptable preferred or approved alternative model), using actual

emission changes associated with any increment-consuming or increment-expanding sources.

The AERMOD model allows for inclusion of these emissions (represented as negative emissions

for the sources expanding increment)31 in the same model run that includes the allowable

increase in emissions from the proposed source and will, therefore, output the net cumulative

concentrations at each receptor established for the modeling domain.32

V.3.2.2 Assessing Secondary PM2.5 Impacts

To assess the impacts from changes in secondary PM2.5 precursor emissions from the new

or modified source, as well as from other increment-consuming sources, the EPA recommends the analysis for each applicable precursor of PM2.5 be conducted collectively based on the two- tiered demonstration approach outlined in EPA’s 2017 Guideline.

31 See discussion about increment expansion in Section V.1.3 of this guidance. 32 The “maximum” cumulative impacts will be output as zero if the cumulative impacts computed in the model are less than zero).

71 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

In recent years, several rules promulgated by the EPA have resulted in control

requirements that have significantly reduced NOX and SO2 precursor emissions affecting ambient

33 PM2.5 concentrations in many areas. This is particularly true in the eastern U.S. As a result, in some cases, the impacts of secondary PM2.5 emissions may be addressed by a demonstration that

provides ambient monitoring data that generally confirms a downward trend in contributions of

precursor emissions occurring after the applicable PM2.5 minor source baseline date (or the major

source baseline date). If it can be confirmed that such secondary emissions reductions have

occurred in a particular baseline area, it may be possible to complete the PM2.5 increments modeling analysis simply by focusing on potential increment consumption associated with direct

PM2.5 emissions. For areas where PM2.5 precursor emission increases from other increment-

consuming sources have occurred since the major or minor source baseline dates, and are, thus,

likely to have added to PM2.5 concentration increases within the baseline area (and, thus,

consume PM2.5 increment), the chemical transport modeling methods (using the emissions input

data applicable to increment analyses) discussed in Section III of this guidance may be

appropriate for estimating the portion of PM2.5 increment consumed due to secondary PM2.5

impacts associated with those increases in precursor emissions.

V.4. Determining Whether a Proposed Source Will Cause or Contribute to an Increment Violation

When a proposed PSD source predicts, through a cumulative impact analysis, that a

33 Such rules include the Finding of Significant Contribution and Rulemaking for Certain States in the Ozone Transport Assessment Group Region for Purposes of Reducing Regional Transport of Ozone (also known as the NOx SIP Call), 63 FR 57356 (October 27, 1998); the Clean Air Interstate Rule (CAIR) Final Rule, 70 FR 25162 (May 12, 2005); CSAPR Final Rule, 76 FR 48208 (August 8, 2011); CSAPR Update for the 2008 Ozone NAAQS (CSAPR Update) Final Rule, 81 FR 74504 (October 26, 2016); and the Mercury and Air Toxics Standards Rule (MATS), 77 FR 9304 (February 16, 2012).

72 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

modeled violation of any PM2.5 increment will occur within the baseline area of concern, a closer

examination of the proposed source’s individual impact(s) at the violating receptor(s) and the

time(s) of violation become important considerations. The EPA’s longstanding policy is that a

proposed PSD source will be considered to cause or contribute to an increment violation if its

impact (primary and secondary) is significant (equal to or greater than the applicable PM2.5 SIL)

at the location and time of the modeled violation.34 Accordingly, a proposed source or

modification generally will not be considered to cause or contribute to an increment violation,

even if it’s modeled impacts equal or exceed the applicable PM2.5 SILs, if it can demonstrate to

the satisfaction of the permitting authority that such significant impacts do not occur at the

location and time of any modeled violation.35 In cases where a proposed PSD source models impacts that equal or exceed the applicable PM2.5 SIL and would cause a new violation of any

PM2.5 increment, it is the EPA’s longstanding policy to allow the PSD applicant to obtain

sufficient offsets, in the form of emissions reductions internally or from another existing source,

to avoid causing the violation at each affected receptor where (and when) a violation is modeled.

In an area where a proposed PSD source would cause or contribute to an existing increment

violation(s), the PSD source cannot be approved for construction unless such existing

violation(s) is entirely corrected at each affected receptor prior to the operation of the proposed

34 See, e.g., 43 FR 26380 at 26401, June 19, 1978; EPA memo titled “Interpretation of ‘Significant Contribution,’” December 16, 1980; EPA memo titled “Air Quality Analysis for Prevention of Significant Deterioration,” July 5, 1988; and more recently, EPA memo titled “Guidance on Significant Impact Levels for Ozone and Fine Particles in the Prevention of Significant Deterioration Permitting Program,” April 17, 2018, Attachment at page 18 (“If the modeled impact is below the recommended SIL value at the violating receptor during the violation, the EPA believes this will be sufficient…to conclude that the source does not cause or contribute to…the predicted violation.”)(Emphasis added). 35 The difficulties associated with combining primary and secondary impacts spatially and temporally were described in Sections III and IV of this guidance. In the case of a PM2.5 increment analysis, as with the PM2.5 NAAQS analysis, the applicant and permitting authority will need to agree upon an approach that best satisfies the required compliance demonstration.

73 Does not represent final Agency action; Draft for public review and comment; 02/10/2020 source.36

36 See, e.g., 43 FR 26380 at 26401, June 19, 1978; 45 FR 52676 at 52678, August 7, 1980; and EPA memo titled “Air Quality Analysis for Prevention of Significant Deterioration,” July 5, 1988. (“…for any increment violation (new or existing) for which the proposed source has a significant impact, the permit should not be approved unless the increment violation is corrected prior to operation of the proposed source.) Note that this policy for the PSD increments differs from the policy for sources that contribute to an existing NAAQS violation, for which the proposed sources needs only compensate for its own adverse impact on the NAAQS violation in accordance with 40 CFR 51.165(b)(3).

74 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

VI. References

U.S. EPA, 1990: New Source Review Workshop Manual: Prevention of Significant Deterioration and Nonattainment Area Permitting – DRAFT. U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711. http://www.epa.gov/ttn/nsr/gen/wkshpman.pdf. U.S. EPA, 1992: Protocol for Determining the Best Performing Model. September 1992. EPA- 454/R-92-025. U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711. NARSTO, 2004. Particulate Matter Assessment for Policy Makers: A NARSTO Assessment. P. McMurry, M. Shepherd, and J. Vickery, eds. Cambridge University Press, Cambridge, England. ISBN 0 52 184287 5. U.S. EPA, 2004: User's Guide to the Building Profile Input Program. EPA-454/R-93-038. U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711. U.S. EPA, 2005. Guideline on Air Quality Models. 40 CFR part 51 Appendix W (70 FR 68218, Nov. 9, 2005). http://www.epa.gov/ttn/scram/guidance/guide/appw_05.pdf. Byun and Schere, 2006: Review of the governing equations, computational algorithms, and other components of the models-3 Community Multiscale Air Quality (CMAQ) modeling system. D. Byun and K. Schere. Applied Mechanics Reviews. 2006; 59, 51-77. Tesche, T., Morris, R., Tonnesen, G., McNally, D., Boylan, J., Brewer, P., 2006. CMAQ/CAMx annual 2002performance evaluation over the eastern US. Atmospheric Environment 40, 4906-4919. U.S. EPA, 2007a: Guidance on the Use of Models and Other Analyses for Demonstrating Attainment of Air Quality Goals for Ozone, PM2.5, and Regional Haze. April 2007. EPA- 454/B-07-002. U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711. http://www.epa.gov/ttn/scram/guidance/guide/final-03-pm-rh- guidance.pdf. U.S. EPA, 2007b: Details on Technical Assessment to Develop Interpollutant Trading Ratios for PM2.5 Offsets. Technical Analysis dated July 23, 2007. U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711. Bergin, M., Russell, A., Odman, T., Cohan, D., and Chameldes, W., 2008. Single Source Impact Analysis Using Three-Dimensional Air Quality Models. Journal of the Air & Waste Management Association. 2008; 58, 1351–1359. Russell, 2008: EPA Supersites Program-related emissions-based particulate matter modeling: Initial applications and advances. A. Russell. Journal of the Air & Waste Management Association. 2008; 58, 289-302. U.S. EPA, 2008: AERSURFACE User’s Guide. EPA-454/B-08-001. U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711. https://www3.epa.gov/ttn/scram/7thconf/aermod/aersurface_userguide.pdf. Civerolo, K., Hogrefe, C., Zalewsky, E., Hao, W., Sistla, G., Lynn, B., Rosenzweig, C., Kinney, P.L., 2010. Evaluation of an 18-year CMAQ simulation: Seasonal variations and long- term temporal changes in sulfate and nitrate. Atmospheric Environment 44, 3745-3752.

75 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

U.S. EPA, 2010: Modeling Procedures for Demonstrating Compliance with PM2.5 NAAQS. Stephen Page Memorandum dated March 23, 2010. U.S. Environmental Protection Agency, Research Triangle Park, North Carolina, 27711. http://www.epa.gov/ttn/scram/Official%20Signed%20Modeling%20Proc%20for%20De mo%20Compli%20w%20PM2.5.pdf. Cai, C., Kelly, J.T., Avise, J.C., Kaduwela, A.P., Stockwell, W.R., 2011. Photochemical modeling in California with two chemical mechanisms: model intercomparison and response to emission reductions. Journal of the Air & Waste Management Association 61, 559-572. Hogrefe, C., Hao, W., Zalewsky, E., Ku, J.-Y., Lynn, B., Rosenzweig, C., Schultz, M., Rast, S., Newchurch, M., Wang, L., 2011. An analysis of long-term regional-scale ozone simulations over the Northeastern United States: variability and trends. Atmospheric Chemistry and Physics 11, 567-582. NACAA, 2011: PM2.5 Modeling Implementation for Projects Subject to National Ambient Air Quality Demonstration Requirements Pursuant to New Source Review. Report from NACAA PM2.5 Modeling Implementation Workgroup dated January 7, 2011. Washington, District of Columbia 20001. http://www.epa.gov/ttn/scram/10thmodconf/review_material/01072011- NACAAPM2.5ModelingWorkgroupReport-FINAL.pdf. U.S. EPA, 2011a: AERSCREEN Released as the EPA Recommended Screening Model. Tyler Fox Memorandum dated April 11, 2011. U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711. http://www.epa.gov/ttn/scram/20110411_AERSCREEN_Release_Memo.pdf. U.S. EPA, 2011b: Revised Policy to Address Reconsideration of Interpollutant Trading Provisions for Fine Particles (PM2.5). Gina McCarthy Memorandum dated July 21, 2011. U.S. Environmental Protection Agency, Washington, District of Columbia 20460. http://www.epa.gov/region7/air/nsr/nsrmemos/pm25trade.pdf. U.S. EPA, 2011c: Additional Clarification Regarding Application of Appendix W Modeling Guidance for the 1-hour NO2 National Ambient Air Quality Standard. Tyler Fox Memorandum dated March 1, 2011. U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711. http://www.epa.gov/ttn/scram/guidance/clarification/Additional_Clarifications_Appendix W_Hourly-NO2-NAAQS_FINAL_03-01-2011.pdf. Yarwood, G., Scorgie, Y., Agapides, N., Tai, E., Karamchandani, P., Duc, H., Trieu, T., Bawden, K., 2011. Ozone Impact Screening Method for New Sources Based on High-order Sensitivity Analysis of CAMx Simulations for NSW Metropolitan Areas. ENVIRON, 2012a. Comparison of Single-Source Air Quality Assessment Techniques for Ozone, PM2.5, other Criteria Pollutants and AQRVs, EPA Contract No: EP-D-07-102. September 2012. 06-20443M6. ENVIRON, 2012b. Evaluation of Chemical Dispersion Models Using Atmospheric Plume Measurements from Field Experiments, EPA Contract No: EP-D-07-102. September 2012. 06-20443M6.

76 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

U.S. EPA, 2012: Sierra Club Petition Grant. Gina McCarthy Administrative Action dated January 4, 2012. U.S. Environmental Protection Agency, Washington, District of Columbia 20460. http://www.epa.gov/ttn/scram/10thmodconf/review_material/Sierra_Club_Petition_OAR- 11-002-1093.pdf. Zhou, W., Cohan, D.S., Pinder, R.W., Neuman, J.A., Holloway, J.S., Peischl, J., Ryerson, T.B., Nowak, J.B., Flocke, F., Zheng, W.G., 2012. Observation and modeling of the evolution of Texas power plant plumes. Atmospheric Chemistry and Physics 12, 455-468. Baker, K.R., Kelly, J.T., 2014. Single source impacts estimated with photochemical model source sensitivity and apportionment approaches. Atmospheric Environment 96, 266-274. Chen, J., Lu, J., Avise, J.C., DaMassa, J.A., Kleeman, M.J., Kaduwela, A.P., 2014. Seasonal modeling of PM 2.5 in California's San Joaquin Valley. Atmospheric Environment 92, 182-190. U.S. EPA, 2014a. Guidance for PM2.5 Modeling. May 20, 2014. Publication No. EPA-454/B- 14-001. Office of Air Quality Planning and Standards, Research Triangle Park, NC. https://www3.epa.gov/ttn/scram/guidance/guide/Guidance_for_PM25_Permit_Modeling. pdf. Baker, K.R., Kotchenruther, R.A., Hudman, R.C., 2015. Estimating Ozone and Secondary PM2.5 Impacts from Hypothetical Single Source Emissions in the Central and Eastern United States. Atmospheric Pollution Research 7, 122-133. Kelly, J.T., Baker, K.R., Napelenok, S.L., Roselle, S.J., 2015. Examining Single-Source Secondary Impacts Estimated from Brute-force, Decoupled Direct Method, and Advanced Plume Treatment Approaches. Atmospheric Environment 111, 10-19. U.S EPA, 2015 AERMINUTE User’s Guide. U.S. Environmental Protection Agency, EPA- 454/b-15-006. Research Triangle Park, North Carolina 27711. https://www3.epa.gov/ttn/scram/7thconf/aermod/aerminute_userguide.pdf. U.S. EPA, 2016a. Guidance on the use of models for assessing the impacts of emissions from single sources on the secondarily formed pollutants ozone and PM2.5. Publication No. EPA 454/R-16-005. Office of Air Quality Planning and Standards, Research Triangle Park, NC. https://www3.epa.gov/ttn/scram/appendix_w/2016/EPA-454_R-16-005.pdf. U.S. EPA, 2016b: AERSCREEN User’s Guide. EPA-454-/B-11-001. U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711. http://www.epa.gov/ttn/scram/models/screen/aerscreen_userguide.pdf. U.S. EPA, 2017a. Guideline on Air Quality Models. 40 CFR part 51 Appendix W (82 FR 5182, Jan. 17, 2017). https://www3.epa.gov/ttn/scram/guidance/guide/appw_17.pdf. U.S. EPA, 2017b. Use of Photochemical Grid Models for Single-Source Ozone and secondary PM2.5 impacts for Permit Program Related Assessments and for NAAQS Attainment Demonstrations for Ozone, PM2.5 and Regional Haze. Tyler Fox Memorandum dated August 4, 2017. U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711. https://www3.epa.gov/ttn/scram/guidance/clarification/20170804- Photochemical_Grid_Model_Clarification_Memo.pdf. Ramboll Environ, 2018. User's Guide Comprehensive Air Quality Model with Extensions version 6. ENVIRON International Corporation, Novato, CA. http://www.camx.com.

77 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

U.S. EPA, 2018a: Guidance on Significant Impact Levels for Ozone and Fine Particles in the Prevention of Significant Deterioration Permitting Program. Office of Air Quality Planning and Standards, Research Triangle Park, NC. https://www.epa.gov/nsr/significant-impact-levels-ozone-and-fine-particles. U.S. EPA, 2018b: User's Guide for the AERMOD Terrain Preprocessor (AERMAP). EPA- 454/B-16-012. U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711. https://www3.epa.gov/ttn/scram/models/aermod/aermap/aermap_userguide_v11105.pdf. U.S. EPA, 2019a. Guidance on the Development of Modeled Emission Rates for Precursors (MERPs) as a Tier 1 Demonstration Tool for Ozone and PM2.5 under the PSD Permitting Program. Publication No. EPA 454/R-19-003. Office of Air Quality Planning and Standards, Research Triangle Park, NC. https://www3.epa.gov/ttn/scram/guidance/guide/EPA-454_R-19-003.pdf. U.S. EPA, 2019b: User's Guide for the AMS/EPA Regulatory Model – AERMOD. EPA-454/B- 19-027. U.S. Environmental Protection Agency, Research Triangle Park, NC 27711. https://www3.epa.gov/ttn/scram/models/aermod/aermod_userguide.pdf. U.S. EPA, 2019c: User’s Guide for the AERMOD Meteorological Preprocessor (AERMET). EPA-454/B-19-028. U.S. Environmental Protection Agency, Research Triangle Park, NC 27711. https://www3.epa.gov/ttn/scram/7thconf/aermod/aermet_userguide.pdf. U.S. EPA, 2019d: AERMOD Implementation Guide. EPA-454/B-19-035. U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711. https://www3.epa.gov/ttn/scram/models/aermod/aermod_implementation_guide.pdf.

78 Does not represent final Agency action; Draft for public review and comment; 02/10/2020

Appendix A: Draft Conceptual Description of O3 and PM2.5 Concentrations in the U.S.

This appendix provides a brief summary of the current O3 and PM2.5 monitoring networks. It also characterizes O3 and PM air quality in terms of their precursor emissions and chemical composition, concentration levels, and spatial and temporal patterns across the nation based on the ambient data and analyses contained in the EPA’s “Integrated Science Assessment for Ozone and Related Photochemical Oxidants,”37 “The Particle Pollution Report,”38 and “Particulate Matter Staff Paper.”39 Such information may be useful for permit applicants in preparing conceptual descriptions, as discussed in this guidance. Permit applicants also encouraged to reference the EPA’s “Air Quality Trends” website at https://www.epa.gov/airtrends for the current O3 and PM2.5 trends and design values.

Conceptual Descriptions of O3

1. O3 Monitoring Networks

To monitor compliance with the NAAQS, state, local, and tribal environmental agencies operate O3 monitoring sites at various locations, depending on the population of the area and typical peak O3 concentrations. In 2015, there were over 1,300 O3 monitors reporting O3 concentration data to EPA. All monitors that currently report O3 concentration data to the EPA use ultraviolet Federal Equivalent Methods (FEMs). Since the highest O3 concentrations tend to be associated with particular seasons for various locations, EPA requires O3 monitoring during specific monitoring seasons which vary by state. The O3 monitoring seasons for each state are listed in Appendix D to 40 CFR part 58.

Figure A-1 shows the locations of all U.S. ambient O3 monitoring sites reporting data to EPA during the 2013-2015 period. The gray dots represent State and Local Ambient Monitoring Stations (SLAMS) which are operated by state and local governments to meet regulatory requirements and provide air quality information to public health agencies. SLAMS monitors make up about 80 percent of the ambient O3 monitoring network in the U.S. The minimum monitoring requirements to meet the SLAMS O3 network design criteria are specified in Appendix D to 40 CFR part 58. The requirements are based on both population and ambient concentration levels for each Metropolitan Statistical Area (MSA). At least one site for each MSA must be designed to record the maximum concentration for that particular area. The blue dots highlight two important subsets of monitoring sites within the SLAMS network: the “National Core” (NCore) network, which consists of about 80 monitoring sites that collect multi- pollutant measurements on a year-round basis, and the “Photochemical Assessment Monitoring

37 U.S. Environmental Protection Agency (2013). Integrated Science Assessment for Ozone and Related Photochemical Oxidants. U.S. Environmental Protection Agency, Research Triangle Park, NC. EPA/600/R-10/076 (2013 ISA), section 3.2.2 found at https://cfpub.epa.gov/ncea/isa/recordisplay.cfm?deid=247492. 38 The Particle Pollution Report: Current Understanding of Air Quality and Emissions through 2003. http://www.epa.gov/airtrends/aqtrnd04/pmreport03/pmcover_2405.pdf#page=1.

39 Particulate Matter Staff Paper: Review completed in 2012. http://www.epa.gov/ttn/naaqs/standards/pm/s_pm_cr_sp.html.

A-1

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

Stations” (PAMS) network, which consists of about 75 monitoring sites that collect summertime measurements of various precursor gases involved O3 formation. The green dots in Figure A-1 represent O3 monitoring sites in the Clean Air Status and Trends Network (CASTNet) which are mostly located in rural areas. There were about 80 CASTNet sites reporting data to EPA in 2015, with sites in the eastern U.S. generally being operated by the EPA, and sites in the western U.S. generally being operated by the National Park Service (NPS).

Finally, the black dots in Figure A-1 represent “Special Purpose” (SPM) monitoring sites, which generally collect data for research studies, public health reporting, or other non-regulatory purposes, and all other O3 monitoring sites which includes monitors operated by tribes, industry, and other federal agencies such as the U.S. Forest Service (USFS).

Figure A-1. Locations of U.S. Ambient O3 Monitoring Sites in 2013-2015

2. O3 Precursor Emissions and Atmospheric Chemistry

O3 is formed by photochemical reactions of precursor gases and is not directly emitted from specific sources. In the stratosphere, O3 occurs naturally and provides protection against harmful solar ultraviolet radiation. In the troposphere, near ground level, O3 forms through atmospheric reactions involving two main classes of precursor pollutants: volatile organic

A-2

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

compounds (VOCs) and nitrogen oxides (NOX). Carbon monoxide (CO) and methane (CH4) are 40 also important for O3 formation over longer time periods.

Emissions of O3 precursor compounds can be divided into anthropogenic and natural source categories, with natural sources further divided into biogenic emissions (from vegetation, microbes, and animals) and abiotic emissions (from biomass burning, lightning, and geogenic sources). Anthropogenic sources, including mobile sources and power plants, account for the majority of NOX and CO emissions. Anthropogenic sources are also important for VOC emissions, though in some locations and at certain times of the year (e.g., southern states during summer), the majority of VOC emissions come from vegetation.41 In practice, the distinction between natural and anthropogenic sources is often unclear, as human activities directly or indirectly affect emissions from what would have been considered natural sources during the preindustrial era. Thus, emissions from plants, animals, and wildfires could be considered either natural or anthropogenic, depending on whether emissions result from agricultural practices, forest management practices, lightning strikes, or other types of events.42

Rather than varying directly with emissions of its precursors, O3 changes in a nonlinear fashion with the concentrations of its precursors. NOX emissions lead to both the formation and destruction of O3, depending on the local quantities of NOX, VOC, radicals, and sunlight. In areas dominated by fresh emissions of NOX, radicals are removed, which lowers the O3 formation rate. In addition, the scavenging of O3 by reaction with NO is called “titration” and is often found in downtown metropolitan areas, especially near busy streets and roads, as well as in power plant plumes. This short-lived titration results in localized areas in which O3 concentrations are suppressed compared to surrounding areas, but which contain NO2 that adds to subsequent O3 formation further downwind. Consequently, O3 response to reductions in NOX emissions is complex and may include O3 decreases at some times and locations and increases of O3 at other times and locations. In areas with relatively low NOX concentrations, such as those found in remote continental areas and rural and suburban areas downwind of urban centers, O3 production typically varies directly with NOX concentrations (e.g., decreases with decreasing NOX emissions). The NOx titration effect is most pronounced in urban core areas which have higher volume of mobile source NOx emissions from vehicles than do the surrounding areas. It should be noted that such locations, which are heavily NOX saturated (or radical limited), tend to have much lower observed O3 concentrations than downwind areas. As a general rule, as NOx emissions reductions occur, one can expect lower O3 values to increase while the higher O3 values would be expected to decrease. NOx reductions are expected to result in a compressed O3 distribution, relative to current conditions.

The formation of O3 from precursor emissions is also affected by meteorological parameters such as the intensity of sunlight and atmospheric mixing. Major episodes of high ground-level O3 concentrations in the eastern United States are associated with slow-moving high pressure systems. High pressure systems during the warmer seasons are associated with the sinking of air, resulting in warm, generally cloudless skies, with light winds. The sinking of air

40 2013 ISA, section 3.2.2. 41 2013 ISA, section 3.2.1. 42 2013 ISA, sections 3.2 and 3.7.1.

A-3

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

results in the development of stable conditions near the surface which inhibit or reduce the vertical mixing of O3 precursors. The combination of inhibited vertical mixing and light winds minimizes the dispersal of pollutants, allowing their concentrations to build up. In addition, in some parts of the United States (e.g., in Los Angeles), mountain barriers limit mixing and result in a higher frequency and duration of days with elevated O3 concentrations. Photochemical activity involving precursors is enhanced during warmer seasons because of the greater availability of sunlight and higher temperatures.43

3. Spatial and Temporal Patterns in Ambient O3 Concentrations

3.1. Diurnal and Seasonal Patterns

Since O3 formation is a photochemical process, it is not surprising that concentration levels have strong diurnal and seasonal patterns. Concentration levels tend to be highest at times when sunlight reaches its highest intensity, namely during the afternoon hours of the late spring and summer months. However, there are other factors at work, such as the influence of biogenic VOC emissions and stratospheric intrusions during the spring months, long-range transport, and traffic patterns which often cause peak NOX emissions to occur during the morning and evening rush hours.

Figure A-2 shows the diurnal pattern in the hourly O3 concentrations based on ambient monitoring data from 2000 to 2015. For each monitoring site, the median (top panel) and 95th percentile (bottom panel) values for each hour of the day were calculated, and each boxplot shows the range of those values for that particular hour across all monitoring sites. The whiskers of each boxplot extend to the 5th and 95th percentiles, the box represents the inter-quartile range, and the centerline represents the median value. The median and 95th percentile values show a consistent pattern in that O3 levels tend to be lowest during the early AM hours, increasing rapidly after sunrise. Concentrations typically reach their peak during the afternoon hours, then decrease at a fairly constant rate throughout the evening and nighttime hours.

Figure A-3 shows the seasonal pattern in the daily maximum 8-hour O3 concentrations based on ambient monitoring data from 2000 to 2015. For each monitoring site, the median (top panel) and 95th percentile (bottom panel) values for each month of the year were calculated, and each boxplot shows the range of those values for that particular month across all monitoring sites. The whiskers of each boxplot extend to the 5th and 95th percentiles, the box represents the inter-quartile range, and the centerline represents the median value. Again, the median and 95th percentile values show a consistent pattern in that O3 levels tend to be highest during the spring and summer months (April to September), and lower during the fall and winter months (October to March).

43 2013 ISA, section 3.2.

A-4

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

th Figure A-2. Distribution of Median and 95 Percentile Hourly O3 Concentrations by Hour of the Day based on 2000-2015 Monitoring Data

A-5

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

th Figure A-3. Distribution of Median and 95 Percentile Daily Maximum 8-hour O3 Concentrations by Month of the Year based on 2000-2015 Monitoring Data

A-6

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

3.2. Spatial Patterns

To determine whether or not the O3 NAAQS has been met at an ambient monitoring site, a statistic commonly referred to as a “design value” must be calculated based on three consecutive years of data collected from that site. The form of the O3 NAAQS design value statistic is the 3-year average of the annual 4th highest daily maximum 8-hour O3 concentration in parts per million (ppm). The O3 NAAQS is met at an ambient monitoring site when the design value is less than or equal to 0.070 ppm. In counties or other geographic areas with multiple monitors, the area-wide design value is defined as the design value at the highest individual monitoring site, and the area is said to have met the NAAQS if all monitors in the area are meeting the NAAQS.

Figure A-4 shows a map of the O3 design values in the U.S. based on data collected during the 2013-2015 period. The highest design values occur in California and near large metropolitan areas such as Dallas, Denver, Houston, New York City, and Phoenix. The lowest design values occur in the Pacific Northwest, the Northern Rockies, the Upper Midwest, and parts of New England and the Southeast. In general, sparsely populated areas tend to have lower design values than more urbanized areas.

Figure A-4. Map of 2013-2015 O3 Design Values in parts per billion (ppb)

A-7

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

3.3. Interannual Variability and Trends

th Figure A-5 shows the national trend in the annual 4 highest daily maximum 8-hour O3 concentration from 2000 to 2015. The solid black line represents the median value for each year based on 838 “trends” sites with complete monitoring records, the dashed lines represent the 25th and 75th percentile values for each year, and the shaded gray area covers the 10th percentile value up to the 90th percentile value for each year. While there is considerable year-to-year variability, overall the trend shows an improvement in O3 air quality over the 15-year period. In fact, the median annual 4th highest value has decreased by 18% since the beginning of the century, and by 24% since 2002.

th Figure A-5. National Trend in the Annual 4 Highest Daily Maximum 8-hour O3 Concentration

Since the national trend is a simple aggregate of the site-level trends, it is also important to look at how these trends vary spatially. Figure A-6 shows a map of the trends at each monitoring site with at least 12 complete years of data from 2000-2015. The magnitude of the trend at each site is computed using the Theil-Sen slope estimator, and the Mann-Kendall statistic is calculated in order to test for statistical significance using a threshold of 0.05. The trend at each monitoring site is classified as Decreasing (p-value < 0.05, slope < 0; blue triangles), No Trend (p-value >= 0.05, white circles), or Increasing (p-value < 0.05, slope > 0; red triangles). The size of each triangle is proportional to the magnitude of the trend at each monitoring site.

A-8

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

Figure A-6 shows that O3 levels have decreased across much of the eastern U.S. as a result of regional control programs such as the NOx SIP Call and the Clean Air Interstate Rule (CAIR). Large reductions have occurred near many urban areas where local control programs have been implemented in addition to the regional controls. In the western U.S., where control programs have been more localized, the reductions have occurred mostly in California and near large urban areas. In other areas most sites have not shown a significant trend, and there are only a handful of sites have shown an increasing trend.

Figure A-6. Map of site-level O3 trends across the U.S. from 2000 to 2015

Variations in meteorological conditions play an important role in determining O3 concentrations. Ozone is more readily formed on warm, sunny days when the air is stagnant. Conversely, O3 generation is more limited when it is cool, rainy, cloudy, or windy. EPA uses a statistical model to adjust for the variability in seasonal average O3 concentrations due to weather conditions to provide a more accurate assessment of the underlying trend in O3 caused by emissions.44 Figure A-7 shows the national trend in the May to September mean of the daily

44 Louise Camalier, William Cox, and Pat Dolwick (2007). The Effects of Meteorology on Ozone in Urban Areas and their use in Assessing Ozone Trends. Atmospheric Environment, Volume 41, Issue 33, October 2007, pages 7127-7137.

A-9

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

maximum 8-hour O3 concentrations from 2000 to 2015 in 111 urban locations. The dotted red line shows the trend in observed O3 concentrations at selected monitoring sites, while the solid blue line shows the underlying O3 trend at those sites after removing the effects of weather. The solid blue lines represent O3 levels anticipated under “typical” weather conditions and serve as a more accurate assessment of the trend in O3 due to changes in precursor emissions.

Figure A-7 shows that after adjusting for the year-to-year variability in meteorology, the overall trend in seasonal average O3 concentrations is much smoother. The adjusted trend clearly shows that the NOX SIP Call program resulted in a sharp decrease in summertime O3 concentrations starting in 2004. The adjusted trend also indicates that O3 levels decreased between 2004 and 2009, followed by a small increase from 2009 to 2012, then continued to decrease after 2012.

Figure A-7. Trend in the May to September mean of the daily maximum 8-hour O3 concentration before (dotted red line) and after (solid blue line) adjusting for year-to-year variability in meteorology.

A-10

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

Conceptual Description of PM2.5

1. PM2.5 Monitoring Networks

1.1. PM Mass Networks

The 1997 promulgation of a fine particulate NAAQS led to deployment of over 1,500 PM2.5 sites (about 1,000 currently in operation) used to determine whether an area complies with the standard. These sites use a Federal Reference Method (FRM) or Federal Equivalent Method (FEM), daily sampling over 24-hours, or every third or sixth day. Nearly 200 additional measurements not meeting FRM or FEM specifications are provided by the chemical speciation sites (Figure A-1). Approximately 450 stations provide indirect measurements of continuous FEM (hourly resolution) PM2.5 mass.

1.2. Interagency Monitoring of Protected Visual Environments (IMPROVE) Program

The IMPROVE network, with over 150 sites, has provided nearly a 20+ year record of major components of PM2.5 (sulfate, nitrate, organic and elemental carbon fractions, and trace metals) in pristine areas of the United States (Figure A-8). IMPROVE is led by the National Park Service; various federal and state agencies support its operations. The primary focus of the network is to track visibility and trends in visibility.

1.3. PM2.5 Chemical Speciation Monitoring

In addition to the IMPROVE network, approximately 200 EPA speciation sites operate in urban areas of the United States to assist PM2.5 assessment efforts. No FRM exists for particulate speciation, which is not directly required to determine attainment, and there are slight differences between monitors and methods used in the Chemical Speciation Network (CSN). However, the network’s coverage (Figure A-8) across urban and rural areas has proved essential for a wide range of research and analysis. The speciation networks typically collect a 24-hour sample every three, and sometimes six, days.

Only a handful of sites provide near continuous speciation data, usually limited to some combination of sulfate, carbon (organic and elemental splits) and nitrate. This enables insight to diurnal patterns for diagnosing various cause-effect phenomena related to emissions characterization, source attribution analysis and model evaluation.

A-11

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

Figure A-8. Locations of chemical speciation sites delineated by program type

2. Composition of PM2.5

Particulate matter (PM) is a highly complex mixture of solid particles and liquid droplets distributed among numerous atmospheric gases which interact with solid and liquid phases. Particles range in size from those smaller than 1 nanometer (10-9 meter) to over 100 microns (1 micron is 10-6 meter) in diameter (for reference, a typical strand of human hair is 70 microns and particles less than about 20 microns generally are not detectable by the human eye). Particles are classified as PM2.5 and PM10-2.5, corresponding to their size (diameter) range in microns and referring to total particle mass under 2.5 and between 2.5 and 10 microns, respectively.

Particles span many sizes and shapes and consist of hundreds of different chemicals. Particles are emitted directly from sources and also are formed through atmospheric chemical reactions and often are referred to as primary and secondary particles, respectively. Particle pollution also varies by time of year and location and is affected by several aspects of weather such as temperature, clouds, humidity, and wind. Further complicating particles is the shifting between solid/liquid and gaseous phases influenced by concentration and meteorology, especially temperature.

Particles are made up of different chemical components. The major components, or species, are carbon, sulfate and nitrate compounds, and crustal materials such as soil and ash (Figure A-9). The different components that make up particle pollution come from specific sources and are often formed in the atmosphere. Particulate matter includes both “primary” PM, which is directly emitted into the air, and “secondary” PM, which forms indirectly from fuel

A-12

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

combustion and other sources. Primary PM consists of carbon (soot) emitted from cars, trucks, heavy equipment, forest fires, and burning waste and crustal material from unpaved roads, stone crushing, construction sites, and metallurgical operations. Secondary PM forms in the atmosphere from gases. Some of these reactions require sunlight and/or water vapor. Secondary PM includes: • Sulfates formed from sulfur dioxide emissions from power plants and industrial facilities; • Nitrates formed from nitrogen oxide emissions from cars, trucks, industrial facilities, and power plants; and • Carbon formed from reactive organic gas emissions from cars, trucks, industrial facilities, forest fires, and biogenic sources such as trees.

In addition, ammonia from sources such as fertilizer and animal feed operations is part of the formation of sulfates and nitrates that exist in the atmosphere as ammonium sulfate and ammonium nitrate. Note that fine particles can be transported long distances by wind and weather and can be found in the air thousands of miles from where they were formed.

The chemical makeup of particles varies across the United States (as shown in Figure A- 10). For example, fine particles in the eastern half of the United States contain more sulfates than those in the West, while fine particles in southern California contain more nitrates than other areas of the country. Organic carbon is a substantial component of fine particle mass everywhere.

Figure A-9. National Average of Source Impacts on Fine Particle Levels

Cars, trucks, heavy equipment, wildfires, wood/waste burning, and biogenics Suspended soil and industrial metallurgical operations

Cars, trucks, industrial combustion, and Industrial combustion and power power generation generation

Source: The Particulate Matter Report, EPA-454-R-04-002, Fall 2004. Carbon reflects both organic carbon and elemental carbon. Organic carbon accounts for automobiles, biogenics, gas-powered off-road, and wildfires. Elemental carbon is mainly from diesel powered sources.

A-13

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

Figure A-10. Annual Average PM2.5 Composition grouped by CBSA: 2013-2015

3. Seasonal and Daily Patterns of PM2.5

Fine particles often have a seasonal pattern. Both daily values and quarterly average of PM2.5 also reveal patterns based on the time of year. Unlike daily O3 levels, which are usually elevated in the summer, daily PM2.5 values at some locations can be high at any time of the year. As shown in Figure A-11, PM2.5 values in the eastern half of the United States are typically higher in the third calendar quarter (July-September) when sulfates are more readily formed from sulfur dioxide (SO2) emissions from power plants in that region and when secondary organic aerosol is more readily formed in the atmosphere. Fine particle concentrations tend to be higher in the first calendar quarter (January through March) in the Midwest in part because fine particle nitrates are more readily formed in cooler weather. PM2.5 values are high during the first (January through March) and fourth calendar quarter (October through December) in many areas of the West, in part because of fine particle nitrates and also due to carbonaceous particles which are directly emitted from wood stove and fireplace use. Average concentration from all locations reporting PM2.5 with valid design values is shown.

A-14

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

-3 Figure A-11. Quarterly Averages of PM2.5 Concentration (μg m ): 2013-2015

The composition of PM2.5 also varies by season and helps explain why mass varies by season. Figure A-12 shows the average composition by season (spring, summer, fall and winter) for PM2.5 data collected during 2013-2015. In the eastern United States, sulfate are high in the spring (March-May) and summer (July-September). Nitrates are most evident in the midwest and western cities where its percentage is moderately high in the winter and fall. Organic carbon (OC) is high throughout the year.

A-15

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

Figure A-12. Quarterly Average PM2.5 Composition grouped by CBSA: 2013-2015

The composition of the highest daily PM2.5 values may be different than that for the annual average. Figure A-13 provides 2013-2015 data PM2.5 composition on high mass days across the United States. Mass is proportioned into six components: sulfates, nitrates, OC, elemental carbon (EC), crustal material, and sea-salt. Except for the southeast (where there is little nitrate in PM2.5), nitrates are slightly higher in the top 10 percent of the PM2.5 days. For the 2013-2015 measurements, the percent of sulfates is currently similar or slightly less on the top 10 percent of the days as compared to the annual averages. The portion of OC appears to be similar on the high days compared to the annual averages, except for the Northern Rockies and Upper Midwest where the high days are influenced by OC from wood stoves/fireplaces and wildfires.

A-16

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

Figure A-13. PM2.5 Composition on 10% highest mass concentration days grouped by CBSA: 2013-2015

A-17

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

This Page Intentionally Left Blank

A-18

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

Appendix B: General Guidance on Use of Dispersion Models for Estimating Primary PM2.5 Concentrations

This appendix provides general guidance on the application of dispersion models for estimating ambient concentrations of PM2.5 associated with direct emissions of primary PM2.5. This guidance is based on and is consistent with the EPA’s Guideline on Air Quality Models, published as Appendix W of 40 CFR part 51, and focuses primarily on the application of AERMOD, the EPA’s preferred dispersion model for most situations. Appendix W is the primary source of information on the regulatory application of air quality models for State Implementation Plan (SIP) revisions for existing sources and for New Source Review (NSR) and Prevention of Significant Deterioration (PSD) programs. There will be applications of dispersion models unique to specific areas, (i.e., there may be areas of the country where it is necessary to model unique specific sources or types of sources). In such cases, there should be consultation with the state or appropriate permitting authority with the appropriate EPA Regional Office modeling contact to discuss how best to model a particular source.

Recently issued EPA guidance of relevance for consideration in modeling for PM2.5 includes: • “Model Clearinghouse Review of Modeling Procedures for Demonstrating Compliance with PM2.5 NAAQS” February 26, 2010 (U.S. EPA, 2010a);

• ”Modeling Procedures for Demonstrating Compliance with PM2.5 NAAQS” March 23, 2010 (U.S. EPA, 2010b); and

• “Transportation Conformity Guidance for Quantitative Hot-spot Analyses in PM2.5 and PM10 Nonattainment and Maintenance Areas” November 2013 (U.S.EPA, 2013a).

The guidance listed above, in addition to other relevant support documents can be found on the SCRAM website at: https://www.epa.gov/scram.

The following sections will refer to the relevant sections of Appendix W and other existing guidance with summaries as necessary. Please refer to those original guidance documents for full discussion and consult with the appropriate EPA Regional Office modeling contact if questions arise about interpretation on modeling techniques and procedures.45

1. Model selection

Preferred air quality models for use in regulatory applications are addressed in Appendix A of the EPA's Guideline on Air Quality Models. If a model is to be used for a particular application, the user should follow the guidance on the preferred model for that application. These models may be used without an area specific formal demonstration of applicability as long as they are used as indicated in each model summary of Appendix A. Further recommendations for the application of these models to specific source problems are found in Appendix W. In

45 A list of EPA Regional Office modeling contacts is available on the SCRAM website at: https://www.epa.gov/scram/air-modeling-regional-contacts.

B-1

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

2005, the EPA promulgated the American Meteorological Society/Environmental Protection Agency Regulatory Model (AERMOD) as the Agency’s preferred near-field dispersion model for a wide range of regulatory applications in all types of terrain based on extensive developmental and performance evaluation. For PSD/NSR modeling under the PM2.5 NAAQS, AERMOD should be used to model primary PM2.5 emissions unless use of an alternative model can be justified (section 3.2, Appendix W).

The AERMOD modeling system includes the following components: • AERMOD: the dispersion model (U.S. EPA, 2019a); • AERMAP: the terrain processor for AERMOD (U.S. EPA, 2018,); and • AERMET: the meteorological data processor for AERMOD (U.S. EPA, 2019b;).

Other components that may be used, depending on the application, are: • BPIPPRIME: the building input processor (U.S. EPA, 2004); • AERSURFACE: the surface characteristics processor for AERMET (U.S. EPA, 2008); • AERSCREEN: a screening version of AERMOD (U.S. EPA, 2016a; U.S. EPA, 2011); and • AERMINUTE: a pre-processor to calculate hourly average winds from ASOS 2-minute observations (U.S. EPA, 2015).

Before running AERMOD, the user should become familiar with the user’s guides associated with the modeling components listed above and the AERMOD Implementation Guide (AIG) (U.S. EPA, 2019c). The AIG lists several recommendations for applications of AERMOD that would be applicable for SIP and PSD permit modeling.

1.2. Receptor grid

The model receptor grid is unique to the particular situation and depends on the size of the modeling domain, the number of modeled sources, and complexity of the terrain. Receptors should be placed in areas that are considered ambient air (i.e., outside of buildings and where the public generally has access) and placed out to a distance such that areas of violation can be detected from the model output to help determine the size of nonattainment areas. Receptor placement should be of sufficient density to provide resolution needed to detect significant gradients in the concentrations with receptors placed closer together near the source to detect local gradients and placed farther apart away from the source. In addition, the user may want to place receptors at key locations such as around facility “fence lines”46 (which define the ambient air boundary for a particular source) or monitor locations (for comparison to monitored

46 It should be noted that the term “fence line” for modeling purposes generally makes reference to a source’s property boundary and may not refer literally to the existence of a fence at such boundary. The EPA’s “ambient air” policy does not mandate that public access to a source’s property be precluded by a fence; other measures that effectively preclude public access may be approved for establishing an ambient air exclusion for PSD modeling purposes.

B-2

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

concentrations for model evaluation purposes). The receptor network should cover the modeling domain. States may already have existing receptor placement strategies in place for regulatory dispersion modeling under NSR/PSD permit programs.

If modeling indicates elevated levels of PM2.5 (near the standard) near the edge of the receptor grid, consideration should be given to expanding the grid or conducting an additional modeling run centered on the area of concern. As noted above, terrain complexity should also be considered when setting up the receptor grid. If complex terrain is included in the model calculations, AERMOD requires that receptor elevations be included in the model inputs. In those cases, the AERMAP terrain processor (U.S. EPA, 2018) should be used to generate the receptor elevations and hill heights. The latest version of AERMAP (version 09040 or later) can process either Digitized Elevation Model (DEM) or National Elevation Data (NED) data files. The AIG recommends the use of NED data since it is more up to date than DEM data, which is no longer updated (Section 4.3 of the AIG).

2. Source inputs

This section provides guidance on source characterization to develop appropriate inputs for dispersion modeling with the AERMOD modeling system. Section 2.1 provides guidance on use of emission, Section 2.2 covers guidance on Good Engineering Practice (GEP) stack heights, Section 2.3 provides details on source configuration and source types, Section 2.4 provides details on urban/rural determination of the sources, and Section 2.5 provides general guidance on source grouping, which may be important for design value calculations.

2.1. Emissions

Consistent with Appendix W, dispersion modeling for the purposes of PSD permitting should be based on the use of continuous operation at maximum allowable emissions or federally enforceable permit limits (see Table 8-2 of Appendix W) for the project source for all applicable averaging periods. Also consistent with past and current guidance, in the absence of maximum allowable emissions or federally enforceable permit limits, potential to emit emissions (i.e., design capacity) should be used. Maximum allowable emissions and continuous operation should also be assumed for nearby sources included in the modeled inventory for the 24-hr PM2.5 NAAQS, while maximum allowable emissions and the actual operating factor averaged over the most recent 2 years should be used for modeled nearby sources for the annual PM2.5 NAAQS.

2.2. Good Engineering Practice (GEP) stack height

Consistent with previous modeling guidance and section 7.2.2.1 of Appendix W, for stacks with heights that are within the limits of Good Engineering Practice (GEP), actual heights should be used in modeling. Under the EPA’s regulations at 40 CFR 51.100, GEP height, Hg, is determined to be the greater of: • 65 m, measured from the ground-level elevation at the base of the stack; • for stacks in existence on January 12, 1979, and for which the owner or operator had obtained all applicable permits or approvals required under 40 CFR parts 51 and 52

B-3

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

Hg=2.5H

provided the owner or operator produces evidence that this equation was actually relied on in designing the stack or establishing an emission limitation to ensure protection against downwash; • for all other stacks,

Hg=H + 1.5L,

where H is the height of the nearby structure(s) measured from the ground-level elevation at the base of the stack and L is the lesser dimension of height or projected width of nearby structure(s); or • the height demonstrated by a fluid model or a field study approved by the EPA or the state/local permitting agency which ensures that the emissions from a stack do not result in excessive concentrations of any air pollutant as a result of atmospheric downwash, wakes, eddy effects created by the source itself, nearby structures or nearby terrain features.

For more details about GEP, see the Guideline for Determination of Good Engineering Practice Stack Height Technical Support Document (U.S. EPA, 1985).

If stack heights exceed GEP, then GEP heights should be used with the individual stack’s other parameters (temperature, diameter, exit velocity). For stacks modeled with actual heights below GEP that may be subject to building downwash influences, building downwash should be considered as this can impact concentrations near the source (section 7.2.2.1(b), Appendix W). If building downwash is being considered, the BPIPPRIME program (U.S. EPA, 2004) should be used to input building parameters for AERMOD.

2.3. Source configurations and source types

An accurate characterization of the modeled facilities is critical for refined dispersion modeling, including accurate stack parameters and physical plant layout. Accurate stack parameters should be determined for the emissions being modeled. Since modeling would be done with maximum allowable or potential emissions levels at each stack, the stack’s parameters such as exit temperature, diameter, and exit velocity should reflect those emissions levels. Accurate locations (i.e., latitude and longitude or Universal Transverse Mercator (UTM) coordinates and datum)47 of the modeled emission sources are also important, as this can affect the impact of an emission source on receptors, determination of stack base elevation, and relative location to any nearby building structures. Not only are accurate stack locations needed, but accurate information for any nearby buildings is important. This information would include

47 Latitudes and longitudes to four decimal places position a stack within 30 feet of its actual location and five decimal places position a stack within three feet of its actual location. Users should use the greatest precision available.

B-4

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

location and orientation relative to stacks and building size parameters (height, and corner coordinates of tiers) as these parameters are input into BPIPPRIME to calculate building parameters for AERMOD. If stack locations and or building information are not accurate, downwash will not be accurately accounted for in AERMOD.

Emission source type characterization within the modeling environment is also important. As stated in the AERMOD User’s Guide (U.S. EPA, 2019a), emissions sources can be characterized as several different source types: POINT sources, capped stacks (POINTCAP), horizontal stacks (POINTHOR), VOLUME sources, OPENPIT sources, LINE sources, buoyant lines sources (BUOYLINE), rectangular AREA sources, circular area sources (AREACIRC), and irregularly shaped area sources (AREAPOLY). While most sources can be characterized as POINT sources, some sources, such as fugitive releases or nonpoint sources (emissions from ports/ships, airports, or smaller point sources with no accurate locations), may be best characterized as VOLUME or AREA type sources. Sources such as flares can be modeled in AERMOD using the parameter input methodology described in Section 2.1.2 of the AERSCREEN User’s Guide (U. S. EPA, 2016a). If questions arise about proper source characterization or typing, users should consult the appropriate EPA Regional Office modeling contact.

2.4. Urban/rural determination

For any dispersion modeling exercise, the urban or rural determination of a source is important in determining the boundary layer characteristics that affect the model’s prediction of downwind concentrations. Figure B-1 gives example maximum 24-hour concentration profiles for a 10 meter stack (Figure B-1a) and a 100 m stack (Figure B-1b) based on urban vs. rural designation. The urban population used for the examples is 100,000. In Figure B-1a, the urban concentration is much higher than the rural concentration for distances less than 750 m from the stack but then drops below the rural concentration beyond 750 m. For the taller stack in Figure B-1b, the urban concentration is much higher than the rural concentration even as distances increase from the source. These profiles show that the urban or rural designation of a source can be quite important.

Determining whether a source is urban or rural can be done using the methodology outlined in section 7.2.1.1 of Appendix W and recommendations outlined in Sections 5.1 through 5.3 in the AIG (U.S. EPA, 2019c). In summary, there are two methods of urban/rural classification described in section 7.2.3 of Appendix W.

The first method of urban determination is a land use method (Appendix W, section 7.2.2.1.1(b)(i)). In the land use method, the user analyzes the land use within a 3 km radius of the source using the meteorological land use scheme described by Auer (1978). Using this methodology, a source is considered urban if the land use types I1 (heavy industrial), I2 (light- moderate industrial), C1 (commercial), R2 (common residential), and R3 (compact residential) are 50 percent or more of the area within the 3 km radius circle. Otherwise, the source is considered a rural source. The second method uses population density and is described in section 7.2.2.1.1(b)(ii) of Appendix W. As with the land use method, a circle of 3 km radius is used. If the population density within the circle is greater than 750 people/km2, then the source is

B-5

Does not represent final Agency action; Draft for public review and comment; 02/10/2020 considered urban. Otherwise, the source is modeled as a rural source. Of the two methods, the land use method is considered more definitive (section 7.2.1.1.b, Appendix W). Caution should be exercised with either classification method. As stated in Section 5.1 of the AIG (U.S. EPA, 2009), when using the land use method, a source may be in an urban area but located close enough to a body of water or other non-urban land use category to result in an erroneous rural classification for the source. The AIG in Section 5.1 cautions users against using the land use scheme on a source by source basis, but advises considering the potential for urban heat island influences across the full modeling domain. When using the population density method, section 7.2.2.1.1(b)(ii)of Appendix W states, “Population density should be used with caution and should not be applied to highly industrialized areas where the population density may be low and thus a rural classification would be indicated, but the area is sufficiently built-up so that the urban land use criteria would be satisfied...” With either method, section 7.2.1.1(f) of Appendix W recommends modeling all sources within an urban complex as urban, even if some sources within the complex would be considered rural using either the land use or population density method.

B-6

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

Figure B-1. Urban (red) and rural (blue) concentration profiles for (a) 10 m buoyant stack release, and (b) 100 m buoyant stack release

B-7

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

Another consideration that may need attention by the user, and is discussed in Section 5.1 of the AIG, relates to tall stacks located within or adjacent to small to moderate size urban areas. In such cases, the stack height or effective plume height for very buoyant sources may extend above the urban boundary layer height. The application of the urban option in AERMOD for these types of sources may artificially limit the plume height. The use of the urban option may not be appropriate for these sources, since the actual plume is likely to be transported over the urban boundary layer. Section 5.1 of the AIG gives details on determining if a tall stack should be modeled as urban or rural based on comparing the stack or effective plume height to the urban boundary layer height. The 100 m stack illustrated in Figure B-1b, may be such an example as the urban boundary layer height for this stack would be 189 m (based on a population of 100,000) and equation 104 of the AERMOD formulation document (Cimorelli, et al., 2004). This equation is:

1  P  4 z = z   iuc iuo  P   o  (B-1) where ziuo is a reference height of 400 m corresponding to a reference population Po of 2,000,000 people.

Given that the stack is a buoyant release, the plume may extend above the urban boundary layer and may be best characterized as a rural source, even if it were near an urban complex. However, beginning with version 15181 of AERMOD, a formulation bug fix was incorporated that modified the treatment of plume rise for urban sources, especially for tall stacks in urban areas. See Section 5.1 of the AIG for more information. Even with the bug fix in AERMOD 15181, exclusion of these elevated sources from application of the urban option would need to be justified on a case-by-case basis in consultation with the appropriate permitting authority.

AERMOD requires the input of urban population when utilizing the urban option. Population can be entered to one or two significant digits (i.e., an urban population of 1,674,365 can be entered as 1,700,000). Users can enter multiple urban areas and populations using the URBANOPT keyword in the runstream file (U.S. EPA, 2019a). If multiple urban areas are entered, AERMOD requires that each urban source be associated with a particular urban area or AERMOD model calculations will abort. Urban populations can be determined by using a method described in Section 5.2 of the AIG (U.S. EPA, 2019c).

2.5. Source groups

In AERMOD, individual emission sources’ concentration results can be combined into groups using the SRCGROUP keyword (Section 3.3.11 of the AERMOD User’s Guide (U.S, EPA, 2019a). The user can automatically calculate a total concentration (from all sources) using the SRCGROUP ALL keyword. For the purposes of design value calculations, source group ALL should be used, especially if all sources in the modeling domain are modeled in one AERMOD run. Design values should be calculated from the total concentrations (all sources and background). Individual source impacts on the total concentration may be necessary to determine the culpability to any NAAQS violations.

B-8

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

3. Meteorological data

This section gives guidance on the selection of meteorological data for input into AERMOD. Much of the guidance from section 8.4 of Appendix W is applicable to SIP and PSD permit modeling and is summarized here. In Section 3.2.1, the use of the tool, AERMINUTE (U.S. EPA, 2015), is introduced. AERMINUTE is an AERMET pre-processor that calculates hourly averaged winds from ASOS (Automated Surface Observing System) 1-minute winds. Section 3.2.4 discusses the use of prognostic meteorological data.

3.1. Surface characteristics and representativeness

The selection of meteorological data that are input into a dispersion model should be considered carefully. The selection of data should be based on spatial and climatological (temporal) representativeness (Appendix W, section 8.4). The representativeness of the data is based on: 1) the proximity of the meteorological monitoring site to the area under consideration, 2) the complexity of terrain, 3) the exposure of the meteorological site, and 4) the period of time during which data are collected. Sources of meteorological data are: National Weather Service (NWS) stations, site-specific or onsite data, and other sources such as universities, Federal Aviation Administration (FAA), military stations, and others. In specific cases, prognostic meteorological data may be appropriate for use and obtained from similar sources. Appendix W addresses spatial representativeness issues in sections 8.4.1.a and 8.4.2.b.

Spatial representativeness of the meteorological data can be adversely affected by large distances between the source and receptors of interest and the complex topographic characteristics of the area (Appendix W, sections 8.4.1.a and 8.4.2.b). If the modeling domain is large enough such that conditions vary drastically across the domain, then the selection of a single station to represent the domain should be carefully considered. Also, care should be taken when selecting a station if the area has complex terrain. While a source and meteorological station may be in close proximity, there may be complex terrain between them such that conditions at the meteorological station may not be representative of the source. An example would be a source located on the windward side of a mountain chain with a meteorological station a few kilometers away on the leeward side of the mountain. Spatial representativeness for off-site data should also be assessed by comparing the surface characteristics (albedo, Bowen ratio, and surface roughness) of the meteorological monitoring site and the analysis area. When processing meteorological data in AERMET (U.S. EPA, 2016c), the surface characteristics of the meteorological site or the prognostic meteorological model output grid cell should be used (section 8.4.2.b of Appendix W and the AERSURFACE User’s Guide (U.S. EPA 2008)). Spatial representativeness should also be addressed for each meteorological variable separately. For example, temperature data from a meteorological station several kilometers from the analysis area may be considered adequately representative, while it may be necessary to collect wind data near the plume height (section 8.4.2.b of Appendix W).

Surface characteristics can be calculated in several ways. For details, see Section 3.1.2 of the AIG (U.S. EPA, 2019c). The EPA has developed a tool, AERSURFACE (U.S. EPA, 2008) to aid in the determination of surface characteristics for observed meteorological data. The current version of AERSURFACE uses the 1992 National Land Cover Data. Note that the use of

B-9

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

AERSURFACE is not a regulatory requirement, but the methodology outlined in Section 3.1.2 of the AIG should be followed unless an alternative method can be justified. For prognostic meteorological output, the surface characteristics of the representative grid cell should be used.

3.2. Meteorological inputs

Appendix W states in section 8.4.2.e that the user should acquire enough meteorological data to ensure that worst-case conditions are adequately represented in the model results. Appendix W states that 5 years of NWS meteorological data, at least 1 year of site-specific data, or at least 3 years of prognostic data should be used and should be adequately representative of the study area. If 1 or more years of site-specific data are available, those data are preferred. While the form of the PM2.5 NAAQS contemplates obtaining 3 years of monitoring data, this does not preempt the use of 5 years of NWS data or at least 1 year of site-specific data in the modeling. The 5-year average based on the use of NWS data, an average across 3 or more years of prognostic data, or an average across 1 or more years of available site specific data, serves as an unbiased estimate of the 3-year average for purposes of modeling demonstrations of compliance with the NAAQS.

3.2.1. NWS data

NWS data are available from the National Climatic Data Center (NCDC) in many formats, with the most common one in recent years being the Integrated Surface Hourly data (ISH). Most available formats can be processed by AERMET. As stated in Section 3.1, when using data from an NWS station alone or in conjunction with site-specific data, the data should be spatially and temporally representative of conditions at the modeled sources. Key points regarding the use of NWS data can be found in the EPA’s March 8, 2013 clarification memo “Use of ASOS meteorological data in AERMOD dispersion modeling” (U.S. EPA, 2013b). The key points are:

• The EPA has previously analyzed the effects of ASOS implementation on dispersion modeling and found that generally AERMOD was less sensitive than ISCST3 to the implementation of ASOS. • The implementation of the ASOS system over the conventional observation system should not preclude the consideration of NWS stations in dispersion modeling. • The EPA has implemented an adjustment factor (0.5 knots) in AERMET to adjust for wind speed truncation in ASOS winds • The EPA has developed the AERMINUTE processor (U.S. EPA, 2015) to process 2- minute ASOS winds and calculate an hourly average for input into AERMET. The use of hourly averaged winds better reflect actual conditions over the hour as opposed to a single 2-minute observation.

3.2.2. Site-specific data

The use of site-specific meteorological data is the best way to achieve spatial

B-10

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

representativeness. AERMET can process a variety of formats and variables for site-specific data. The use of site-specific data for regulatory applications is discussed in detail in section 8.4.4 of Appendix W. Due to the range of data that can be collected onsite and the range of formats of data input to AERMET, the user should consult Appendix W, the AERMET User’s Guide (U.S. EPA, 2016c), and Meteorological Monitoring Guidance for Regulatory Modeling Applications (U.S. EPA, 2000). Also, when processing site-specific data for an urban application, Section 3.3 of the AERMOD Implementation Guide offers recommendations for data processing. In summary, the guide recommends that site-specific turbulence measurements should not be used when applying AERMOD’s urban option in order to avoid double counting the effects of enhanced turbulence due to the urban heat island.

3.2.3. Upper air data

AERMET requires full upper air soundings to calculate the convective mixing height. For AERMOD applications in the U.S., the early morning sounding, usually the 1200 UTC (Universal Time Coordinate) sounding, is typically used for this purpose. Upper air soundings can be obtained from the Radiosonde Data of North America CD for the period 1946-1997. Upper air soundings for 1994 through the present are also available for free download from the Radiosonde Database Access website. Users should choose all levels or mandatory and significant pressure levels48 when selecting upper air data. Selecting mandatory levels only would not be adequate for input into AERMET as the use of just mandatory levels would not provide an adequate characterization of the potential temperature profile.

3.2.3. Prognostic data

In specific situations where it is infeasible or cost prohibitive to collect adequately representative site-specific data or there is not a representative NWS or comparable meteorological station available, it may be appropriate to use prognostic meteorological data, if deemed adequately representative. However, if prognostic data are not representative of the transport and dispersion conditions in the area of concern, the collection of site-specific data is necessary (section 8.4.5.1 of Appendix W). To facilitate the use of prognostic meteorological data, EPA has developed a processor, Mesoscale Model Interface Program, MMIF (Environ, 2015), to process MM5 (Mesoscale Model 5) or WRF (Weather Research Forecast) model data for input to various models including AERMOD. MMIF can process data for input to AERMET or AERMOD for a single grid cell or multiple grid cells. For regulatory applications, MMIF should be run to create inputs for AERMET input as described in section 8.4.5.1.b of Appendix W and MMIF guidance (U.S. EPA, 2016b). Specific guidance on running MMIF for AERMOD applications can be found in U.S. EPA, 2016b.

4. Running AERMOD and implications for design value calculations

Recent enhancements to AERMOD include options to aid in the calculation of design

48 By international convention, mandatory levels are in millibars: 1,000, 850, 700, 500, 400, 300, 200, 150, 100, 50, 30, 20, 10, 7 5, 3, 2, and 1. Significant levels may vary depending on the meteorological conditions at the upper-air station.

B-11

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

values for comparison with the PM2.5 NAAQS and to aid in determining whether emissions from the project source caused or contributed to any modeled violations. These enhancements include: • The MAXDCONT option, which shows the impact of each user-specified source group to the high ranked values for a specified target source group paired in time and space. The user can specify a range of ranks to analyze or specify an upper bound rank, i.e., 8th th highest, corresponding to the 98 percentile for the 24-hour PM2.5 NAAQS, and a lower threshold concentration value, such as the NAAQS for the target source group. The model will process each rank within the range specified, but will stop after the first rank (in descending order of concentration) that is below the threshold value if specified by the user. A warning message will be generated if the threshold is not reached within the range of ranks analyzed (based on the range of ranks specified on the RECTABLE keyword). This option may be needed to aid in determining which sources should be considered for controls.

For more details about the enhancements, see the AERMOD User’s Guide (U. S. EPA, 2019a).

Ideally, all explicitly modeled sources, receptors, and background should be modeled in one AERMOD run for all modeled years. In this case, one of the above output options can be used in AERMOD to calculate design values for comparison to the NAAQS and determine the area’s attainment status and/or inform attainment/nonattainment boundaries. The use of these options in AERMOD allows AERMOD to internally calculate concentration metrics that can be used to calculate design values and, therefore, lessen the need for large output files, i.e., hourly POSTFILES.

However, there may be situations where a single AERMOD run with all explicitly modeled sources is not possible. These situations often arise due to runtime or storage space considerations during the AERMOD modeling. Sometimes separate AERMOD runs are done for each facility or group of facilities, or by year, or the receptor network is divided into separate sub-networks. In some types of these situations, the MAXDCONT output option may not be an option for design value calculations, especially if all sources are not included in a single run. If the user wishes to utilize one of the three output options, then care should be taken in developing the model inputs to ensure accurate design value calculations.

Situations that would effectively preclude the use of the MAXDCONT option to calculate meaningful AERMOD design value calculations include the following examples: • Separate AERMOD runs for each source or groups of sources. o SIP modeling includes 10 facilities for 5 years of NWS data and each facility is modeled for 5 years in a separate AERMOD run, resulting in ten separate AERMOD runs. • Separate AERMOD runs for each source and each modeled year. o 10 facilities are modeled for 5 years of NWS data. Each facility is modeled separately for each year, resulting in fifty individual AERMOD runs.

In the two situations listed above, the MAXDCONT option would not be useful as the

B-12

Does not represent final Agency action; Draft for public review and comment; 02/10/2020 different AERMOD runs do not include a total concentration with impacts from all facilities. In these situations, the use of 24-hour POSTFILES, which can be quite large, and external post- processing would be needed to calculate design values.

Situations in which the MAXDCONT options may be used but may necessitate some external post-processing afterwards to calculate a design value include: • The receptor network is divided into sections and an AERMOD run, with all sources and years, is made for each sub-network.

o A receptor network of 1,000 receptors is divided into four 250 receptor sub- networks. 10 facilities are modeled with 5 years of NWS data in one AERMOD run for each receptor network, resulting in four AERMOD runs. After the AERMOD runs are complete, the MAXDCONT results for each network can be re-combined into the larger network. • All sources and receptors are modeled in an AERMOD run for each year. • Ten facilities are modeled with 5 years of NWS data. All facilities are modeled with all receptors for each year individually, resulting in five AERMOD runs. MAXDCONT output can be used and post-processed to generate the necessary design value concentrations. The receptor network is divided and each year is modeled separately for each sub-network with all sources. • Ten facilities are modeled with 5 years of NWS data for 1,000 receptors. The receptor network is divided into four 250 receptor networks. For each sub-network, all ten facilities are modeled for each year separately, resulting in twenty AERMOD runs. MAXDCONT output can be used and post-processed to generate the necessary design value concentrations.

B-13

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

5. References

Auer, Jr., A.H., 1978: Correlation of Land Use and Cover with Meteorological Anomalies. Journal of Applied Meteorology, 17(5), 636-643. Brode, R., K. Wesson, J. Thurman, and C. Tillerson, 2008: AERMOD Sensitivity to the Choice of Surface Characteristics, Paper 811, Air And Waste Management Association Annual Conference. Cimorelli, A. J., S. G. Perry, A. Venkatram, J. C. Weil, R. J. Paine, R. B. Wilson, R. F. Lee, W. D. Peters, R. W. Brode, and J. O. Paumier, 2004. AERMOD: Description of Model Formulation, EPA-454/R-03-004. U.S. Environmental Protection Agency, Research Triangle Park, NC. http://www.epa.gov/ttn/scram/7thconf/aermod/aermod_mfd.pdf. U.S. EPA, 1985: Guideline for Determination of Good Engineering Practice Stack Height (Technical Support Document for the Stack Height Regulations), Revised. EPA-450/4- 80-023R. U.S. Environmental Protection Agency, Research Triangle Park, NC 27711. http://www.epa.gov/ttn/scram/guidance/guide/gep.pdf. U.S. EPA, 1992: Screening Procedures for Estimating the Air Quality Impact of Stationary Sources. EPA-454/R-92-019. U.S. Environmental Protection Agency, Research Triangle Park, NC 27711. http://www.epa.gov/ttn/scram/guidance/guide/scrng.wpd. U.S. EPA, 1994: SO2 Guideline Document. EPA-452/R-95-008. U.S. Environmental Protection Agency, Research Triangle Park, NC 27711. U.S. EPA, 2000: Meteorological Monitoring Guidance for Regulatory Modeling Applications. EPA-454/R-99-005. U.S. Environmental Protection Agency, Research Triangle Park, NC 27711. http://www.epa.gov/ttn/scram/guidance/met/mmgrma.pdf. U.S. EPA, 2004: User's Guide to the Building Profile Input Program. EPA-454/R-93-038. U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711. U.S. EPA, 2008: AERSURFACE User’s Guide. EPA-454/B-08-001. U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711. http://www.epa.gov/ttn/scram/7thconf/aermod/aersurface_userguide.pdf. U.S. EPA, 2010a: Model Clearinghouse Review of Modeling Procedures for Demonstrating Compliance with PM2.5 NAAQS. Tyler Fox Memorandum dated February 26, 2010. U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711. http://www.epa.gov/ttn/scram/guidance/mch/new_mch/MCmemo_Region6_PM25_NAA QS_Compliance.pdf. U.S. EPA, 2010b: Modeling Procedures for Demonstrating Compliance with PM2.5 NAAQS. Stephen Page Memorandum dated March 23, 2010. U.S. Environmental Protection Agency, Research Triangle Park, North Carolina, 27711. http://www.epa.gov/ttn/scram/Official%20Signed%20Modeling%20Proc%20for%20De mo%20Compli%20w%20PM2.5.pdf. U.S. EPA, 2011: AERSCREEN Released as the EPA Recommended Screening Model. Tyler Fox Memorandum dated April 11, 2011. U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711. http://www.epa.gov/ttn/scram/20110411_AERSCREEN_Release_Memo.pdf.

B-14

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

U.S. EPA, 2013a: Transportation Conformity Guidance for Quantitative Hot-spot Analyses in PM2.5 and PM10 Nonattainment and Maintenance Areas. November 2013. EPA-420-B- 10-040. U.S. Environmental Protection Agency, Ann Arbor, Michigan 48105. http://www.epa.gov/oms/stateresources/transconf/policy/420b13053-sec.pdf and http://www.epa.gov/oms/stateresources/transconf/policy/420b13053-appx.pdf. U.S. EPA, 2013b: Use of ASOS meteorological data in AERMOD dispersion modeling. Tyler Fox Memorandum dated March 8, 2013. U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711. http://www.epa.gov/ttn/scram/guidance/clarification/20130308_Met_Data_Clarification. pdf. U.S EPA, 2015 AERMINUTE User’s Guide. U.S. Environmental Protection Agency, EPA- 454/b-15-006. Research Triangle Park, North Carolina 27711. https://www3.epa.gov/ttn/scram/7thconf/aermod/aerminute_userguide.pdf. U.S. EPA, 2016a: AERSCREEN User’s Guide. EPA-454-/B-11-001. U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711. http://www.epa.gov/ttn/scram/models/screen/aerscreen_userguide.pdf. U.S. EPA, 2016b: Guidance on the Use of the Mesoscale Model Interface Program (MMIF) for AERMOD Applications. EPA-454/B-16-003. U.S. Environmental Protection Agency, Research Triangle Park, NC 27711. https://www3.epa.gov/ttn/scram/appendix_w/2016/MMIF_Guidance.pdf. U.S. EPA, 2017. Guideline on Air Quality Models. 40 CFR part 51 Appendix W. https://www3.epa.gov/ttn/scram/guidance/guide/appw_17.pdf. U.S. EPA, 2018: User's Guide for the AERMOD Terrain Preprocessor (AERMAP). EPA-454/B- 18-004. U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711. https://www3.epa.gov/ttn/scram/models/aermod/aermap/aermap_userguide_v18081.pdf. U.S. EPA, 2019a: User's Guide for the AMS/EPA Regulatory Model – AERMOD. EPA-454/B- 19-027. U.S. Environmental Protection Agency, Research Triangle Park, NC 27711. https://www3.epa.gov/ttn/scram/models/aermod/aermod_userguide.pdf. U.S. EPA, 2019b: User’s Guide for the AERMOD Meteorological Preprocessor (AERMET). EPA-454/B-19-028. U.S. Environmental Protection Agency, Research Triangle Park, NC 27711. https://www3.epa.gov/ttn/scram/7thconf/aermod/aermet_userguide.pdf. U.S. EPA, 2019c: AERMOD Implementation Guide. EPA-454/B-19-035. U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711. https://www3.epa.gov/ttn/scram/models/aermod/aermod_implementation_guide.pdf.

B-15

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

This Page Intentionally Left Blank

B-16

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

Appendix C: Example of a Tier 1 Demonstration of the Potential for O3 and Secondary PM2.5 Formation

In 2018, a permit applicant, the Tennessee Valley Authority (TVA) Gleason Combustion Turbine Plant, worked closely with the Tennessee Department of Environment and Conservation (TDEC) and EPA Region 4 to develop a compliance demonstration for a major facility modification, including the use of a Tier 1 assessment of O3 and secondary PM2.5 impacts. This Tier 1 assessment was based on the application of Modeled Emission Rates for Precursors (MERPs) and related modeling guidance released by the EPA. In April 2018, the TDEC published state modeling guidance that can be used by PSD applicants in Tennessee that largely restated the technical aspects of the guidance presented in the EPA’s 2016 Draft MERPs Guidance.49 In support of the 2016 Draft MERPs Guidance, the EPA performed photochemical modeling for four hypothetical sources from within Tennessee or in close proximity to Tennessee (Shelby County, TN, Giles County, TN, Barren County, KY and Ashe County, NC), that can be used to represent the O3 and secondary PM2.5 pollutant formation from other large sources in Tennessee (Figure 1).

FIGURE 1

49 The EPA released a draft version of the “Guidance on the Development of Modeled Emission Rates for Precursors (MERPs) as a Tier 1 Demonstration Tool for Ozone and PM2.5 under the PSD Permitting Program” on December 2, 2016, for public review and comment. Based on the feedback gained from this draft, the EPA released a non-draft or final version of the “MERPs Guidance” on April 30, 2019. The information in the 2016 draft MERPs Guidance from which the TDEC based their April 2018 modeling guidance did not substantively change and is representative of information contained in the current 2019 final version of the MERPs Guidance. The 2019 final MERPs Guidance is available at: https://www3.epa.gov/ttn/scram/guidance/guide/EPA-454_R-19-003.pdf.

C-1

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

Assessment of PM2.5

Based on information in the EPA’s 2016 Draft MERPs Guidance, the lowest, most conservative MERPs from these four hypothetical source locations were established in the TDEC state modeling guidance as the default MERPs that can be used throughout Tennessee without the need for further justification (Table 1). The TVA used these default MERPs to assess secondary PM2.5 impacts for the proposed modification at the Gleason facility.

TABLE 1 Default MERPs for Use in TN PSD Applications [1,2] Precursor MERPs for 8-hr O3 MERPs for Daily MERPs for (tons/yr) PM2.5 (tons/yr) Annual PM2.5 (tons/yr) NOx 156 4,000 7,407 SO2 - 667 6,061 VOC 1,339 - - Notes: 1. EPA, 2016 2. TDEC, 2018.

The combined primary and secondary impacts of PM2.5 for the source impact analysis were assessed using the highest (AERMOD) modeled primary PM2.5 concentration (HMC), the Class II SIL, precursor emissions, and the default MERPs. If the sum of the ratios in Equation 4.1 below is less than 1, then the combined PM2.5 impacts are below the PM2.5 SIL, an adequate compliance demonstration has been performed, and no additional analyses are necessary.

The following equation was used for this assessment:

Where:

HMC = Highest modeled primary PM2.5 impact using AERMOD and project related 3 PM2.5 emissions (µg/m ) SIL = Significant Impact Level (µg/m3) NOx_Em = Project related NOx Emissions (tons per year – tpy) NOx MERP = From Table 1 (tpy) SO2_Em = Project related SO2 Emissions (tpy) SO2_MERP = From Table 1 (tpy)

TVA’s 24-hour and annual PM2.5 inputs to Equation 4.1 are provided in Table 2 below, and the resulting impacts are calculated in Equation 4-2 and Equation 4-3 below, respectively.

C-2

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

TABLE 2 [1,2] Secondary PM2.5 Inputs for the SILs in Class II Areas Secondary PM2.5 Impacts 24-hr Annual Average Average Highest Modeled Primary 0.49 0.053 3 [3] PM2.5 Concentration (μg/m ) SILs for the NAAQS and PSD Increments in Class II areas (μg/m3) [4] 1.2 0.2 [5] GCC NOx Emissions (tons/yr) 2,270 2,270 [1] Default NOx MERPs 4,000 7,407 [5] GCC SO2 Emissions (tons/yr) 14.2 14.2 [1] Default SO2 MERPs 667 6,061 Notes: 1. EPA, 2016 and TDEC, 2018. 2. Calculations taken from “GCC_SecPM25_O3_calcs_20180912.xlsx” provided on optical disc. 3. PM2.5 modeling results (Table 4-9). 4. SILs for the NAAQS in Class I and Class II areas and for PSD increments in Class II areas. Based on the April 17, 2018 EPA memo, Guidance on Significant Impact Levels for Ozone and Fine Particles in the Prevention of Significant Deterioration Permitting Program. 5. Emissions taken from Table 3 in “Gleasn PSD Modemssn SA 20180831.xlsx” (provided by TVA to TDEC on optical disc).

Combined Impacts for 24-hour PM2.5 for the SIL in Class II Areas:

Combined Impacts for Annual PM2.5 for the SIL in Class II Areas:

Both 24-hour and annual PM2.5 impacts were less than 1, which indicated that PM2.5 impacts were expected to be below the Class II SILs for the NAAQS and PSD increments. This indicated that emissions from TVA Gleason would not cause or contribute to a violation of the PM2.5 NAAQS in Class II areas.

C-3

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

Assessment of O3

A somewhat more refined analysis was performed to assess the impacts of the proposed project on O3 concentrations in the area around the facility. Application of the TDEC default NOX and VOC MERPs for O3 shown in Table 1 above indicated that O3 impacts would be greater than the 8-hour O3 SIL of 1 ppb and that a cumulative O3 assessment would be necessary to demonstrate whether the facility modification would cause or contribute to a violation of a the O3 NAAQS.

The O3 assessment first examined ambient O3 concentrations in the region surrounding TVA Gleason (GCC). There are no ambient O3 monitors in the immediate vicinity of GCC, but there are six monitors within 150 km of the facility (Figure 2 and Tables 3 and 4). The Cadiz, KY, monitor was selected as the most representative background site due to its proximity to GCC, its comparable levels of precursor emissions in the county, and it has the largest measurement scale indicating it is representative of regional air quality. The three-year average (2015- 2017) of the fourth-highest 8-hour O3 concentration was 61 ppb, well below the 70 ppb NAAQS.

FIGURE 2

C-4

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

TABLE 3

TABLE 4

Site Name Site ID 3 Year Avg. 4th High 8-Hr Ozone Conc. (ppb) Jackson Purchase 21-145-1024 62 Cadiz 21-221-9991 61 Smithland 21-139-0003 64 Fairview 47-187-0106 60 Hopkinsville 21-047-0006 61 Edmund Orgill Park 47-157-1004 65

As previously discussed, in April 2018, TDEC published modeling guidance on the use of EPA’s MERPs in Tennessee (TDEC, 2018) that identified four hypothetical sites, located in Shelby County, TN, Giles County, TN, Barren County, KY and Ashe County, NC, to represent Tennessee sources (Figure 1). Precursor emissions in these four counties were compared to Weakley County, where GCC is located. Weakley County precursor emissions are comparable to emissions in the three rural counties (Giles, Barren and Ashe) and are much lower than Shelby County which is urban (Table 5). Ashe County is much further from GCC and is located in mountainous terrain, unlike the relatively flat terrain around GCC. Both Giles County and Barren County have similar terrain features to Weakley County. NOX MERPs at these two sites are also lower than in Shelby County and Ashe County, which makes the analysis more conservative as ozone impacts from GCC are dominated by NOX emissions.

C-5

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

TABLE 5

For the two most representative hypothetical sources selected, as part of EPA’s MERPs Guidance, the EPA performed photochemical modeling for two hypothetical source heights (low and high stack releases) and three hypothetical emission rates (500, 1000, and 3000 tons per year). As can be seen in Table 6 below, predicted O3 impacts are nonlinear with respect to precursor emissions. At these hypothetical sources, the amount of O3 formed from 3,000 tons of NOX is substantially less than six times the amount formed from 500 tons of NOX on a per ton basis, so using a MERP based on 500 tons of NOX would significantly over-estimate the O3 impacts from GCC. Therefore, this analysis used the most conservative MERPs based on emission rates most similar to emissions from GCC (hypothetical source emissions of 3,000 tons per year for NOX and 500 tons per year for VOCs) at the two most representative sites (Giles County and Barren County) (Table 7).

TABLE 6

PRECURSOR POLL State County FIPS TPY Stack CONC MERP Ht NOX OZONE Kentucky Barren 21009 500 10 2.908 172

NOX OZONE Kentucky Barren 21009 500 90 2.946 170

NOX OZONE Kentucky Barren 21009 1000 90 5.026 199

NOX OZONE Kentucky Barren 21009 3000 90 10.687 281

NOX OZONE Tennessee Giles 47055 500 10 2.616 191

NOX OZONE Tennessee Giles 47055 500 90 3.208 156

NOX OZONE Tennessee Giles 47055 1000 90 5.387 186

NOX OZONE Tennessee Giles 47055 3000 90 10.356 290 GCC Project Emissions are 2,270 for NOx and 158 tpy for VOC.

C-6

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

TABLE 7

The O3 impacts for the source impact assessment were calculated as the sum of the ratio of precursor emissions to the MERPs. If the sum of the ratios is less than 1, then O3 impacts are below the O3 SIL and no cumulative analysis is necessary.

Where:

NOx_Em = Project related NOx Emissions (tons per year – tpy) NOx MERP = From Table 7 (tpy) VOC_Em = Project related VOC Emissions (tpy) VOC_MERP = From Table 7 (tpy)

GCC’s ozone inputs to Equation 4.4 are provided in Table 8, and the resulting impacts are shown in Equation 4.5.

C-7

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

TABLE 8

According to Equation 4.5, the sum of the ratios was greater than 1, and O3 impacts are above the SIL. Therefore, a cumulative O3 analysis was necessary and performed, which added background O3 and compared the combined impacts to the NAAQS as shown in Equation 4.6.

Where:

Background Ozone = 2015-2017 8-hour ozone design value (ppb) for Cadiz monitor NOx_Em = Project related NOx Emissions (tons per year – tpy) NOx MERP = From Table 7 (tpy) VOC_Em = Project related VOC Emissions (tpy) VOC_MERP = From Table 7 (tpy) SIL = 1 ppb ozone NAAQS = 8-hour ozone NAAQS (70 ppb)

Cumulative O3 impacts from GCC are shown below. Using the 3-year 8-hour ozone design value of 61 ppb from Cadiz, KY, the ratios defined in Equation 4.5, and the O3 SIL of 1 ppb, the cumulative O3 impacts did not exceed the NAAQS. This indicated that emissions from GCC would not cause or contribute to a violation of the O3 NAAQS.

61 + [(2,270 ÷ 281) + (158 ÷ 8,333)] * 1 ppb = 69.1 ppb

61 + [8.08 + .02] * 1 = 69.1 ppb

61 + 8.1 = 69.1 ppb

C-8

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

Appendix D: Example of the background monitoring data calculations for a Second Level 24-hour modeling analysis

This appendix provides an illustrative example of the calculations and data sorting recommendations for the background monitoring data to be used in a Second Level 24-hour PM2.5 modeling analysis. In this example, it was determined through discussion and coordination with the appropriate permitting authority that the impacts from the project source’s primary PM2.5 emissions were most prominent during the cool season and were not temporally correlated with background PM2.5 levels that were typical highest during the warm season. So, combining the modeled and monitored levels through a First Level 24-hour PM2.5 modeling analysis was determined to be potentially overly conservative. Extending the compliance demonstration to a Second Level analysis allows for a more refined and appropriate assessment of the cumulative impacts on the primary PM2.5 emissions in this particular situation.

The example provided is from an idealized Federal Reference Method (FRM) PM2.5 monitoring site that operates on a daily (1-in-1 day) frequency with 100% data completeness. In this case, the annual 98th percentile concentration is the 8th highest concentration of the year. In most cases, the FRM monitoring site will likely operate on a 1-and-3 day frequency and will also likely have missing data due to monitor maintenance or collected data not meeting all of the quality assurance criteria. Please reference Appendix N to 40 CFR part 50 to determine the appropriate 98th percentile rank of the monitored data based on the monitor sampling frequency and valid number of days sampled during each year.

The appropriate seasonal (or quarterly) background concentrations to be included as inputs to the AERMOD model per a Second Level 24-hour PM2.5 modeling analysis are as follows:

• Step 1 – Start with the most recent 3-years of representative background PM2.5 ambient monitoring data that are being used to develop the monitored background PM2.5 design value. In this example, the 3-years of 2008 to 2010 are being used to determine the monitored design value.

th • Step 2 – For each year, determine the appropriate rank for the daily 98 percentile PM2.5 concentration. Again, this idealized example is from a 1-in-1 day monitor with 100% data th th completeness. So, the 8 highest concentration of each year is the 98 percentile PM2.5 th concentration. The 98 percentile PM2.5 concentration for 2008 is highlighted in Table E- 1. The full concentration data from 2009 and 2010 are not shown across the steps in this Appendix for simplicity but would be similar to that of 2008.

• Step 3 – Remove from further consideration in this analysis the PM2.5 concentrations th from each year that are greater than the 98 percentile PM2.5 concentration. In the case presented for a 1-in-1 day monitor, the top 7 concentrations are removed. If the monitor were a 1-in-3 day monitor, only the top 2 concentrations would be removed. The resultant dataset after the top 7 concentrations have been removed from further consideration in this analysis for 2008 is presented in Table E-2.

D-1

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

• Step 4 – For each year, divide the resultant annual dataset of the monitored data equal to th or less than the 98 percentile PM2.5 concentration into each season (or quarter). For 2008, the seasonal subsets are presented in Table E-3.

• Step 5 – Determine the maximum PM2.5 concentration from each of the seasonal (or quarterly) subsets created in Step 4 for each year. The maximum PM2.5 concentration from each season for 2008 is highlighted in Table E-3.

• Step 6 – Average the seasonal (or quarterly) maximums from Step 5 across the three years of monitoring data to create the four seasonal background PM2.5 concentrations to be included as inputs to the AERMOD model. These averages for the 2008 to 2010 dataset used in this example are presented in Table E-4. As noted above, the full concentration data from 2009 and 2010 are not shown across the steps in this Appendix for simplicity, but the seasonal maximums from 2009 and 2010 presented in Table E-4 were determined by following the previous five steps similar to that of 2008.

D-2

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

Table E-1. 2008 Daily PM2.5 Concentrations Date Conc. Date Conc. Date Conc. Date Conc. Date Conc. Date Conc. Date Conc. Date Conc. 1-Jan 10.4 16-Feb 15.1 2-Apr 10.5 18-May 11.1 3-Jul 17.1 18-Aug 18.7 3-Oct 12.3 18-Nov 4.4 2-Jan 5.4 17-Feb 11.8 3-Apr 8.2 19-May 7.7 4-Jul 19.8 19-Aug 21.5 4-Oct 19.5 19-Nov 8.2 3-Jan 10.0 18-Feb 3.4 4-Apr 9.7 20-May 13.6 5-Jul 14.3 20-Aug 20.1 5-Oct 23.7 20-Nov 11.1 4-Jan 16.4 19-Feb 4.5 5-Apr 6.9 21-May 12.1 6-Jul 11.5 21-Aug 18.4 6-Oct 19.8 21-Nov 5.3 5-Jan 11.2 20-Feb 4.8 6-Apr 6.3 22-May 10.0 7-Jul 14.3 22-Aug 16.7 7-Oct 21.7 22-Nov 8.9 6-Jan 11.1 21-Feb 11.9 7-Apr 7.9 23-May 13.3 8-Jul 12.2 23-Aug 13.8 8-Oct 12.2 23-Nov 14.0 7-Jan 10.2 22-Feb 20.1 8-Apr 9.8 24-May 11.2 9-Jul 11.1 24-Aug 19.0 9-Oct 5.1 24-Nov 12.7 8-Jan 11.4 23-Feb 11.4 9-Apr 16.5 25-May 17.7 10-Jul 9.7 25-Aug 17.6 10-Oct 10.2 25-Nov 9.7 9-Jan 8.1 24-Feb 19.3 10-Apr 13.3 26-May 14.2 11-Jul 16.4 26-Aug 15.4 11-Oct 10.7 26-Nov 12.8 10-Jan 9.4 25-Feb 18.2 11-Apr 11.0 27-May 15.4 12-Jul 21.5 27-Aug 12.6 12-Oct 5.6 27-Nov 16.6 11-Jan 5.7 26-Feb 12.8 12-Apr 8.8 28-May 13.9 13-Jul 25.1 28-Aug 12.1 13-Oct 5.9 28-Nov 17.2 12-Jan 8.9 27-Feb 5.5 13-Apr 6.3 29-May 9.3 14-Jul 11.7 29-Aug 10.1 14-Oct 9.7 29-Nov 16.6 13-Jan 18.1 28-Feb 9.7 14-Apr 5.1 30-May 14.5 15-Jul 18.9 30-Aug 17.2 15-Oct 12.8 30-Nov 4.5 14-Jan 11.0 29-Feb 12.1 15-Apr 7.9 31-May 20.5 16-Jul 28.9 31-Aug 19.9 16-Oct 16.4 1-Dec 7.5 15-Jan 11.8 1-Mar 9.6 16-Apr 8.2 1-Jun 15.3 17-Jul 27.6 1-Sep 19.4 17-Oct 12.0 2-Dec 10.6 16-Jan 10.7 2-Mar 5.6 17-Apr 14.7 2-Jun 11.5 18-Jul 12.8 2-Sep 18.2 18-Oct 7.9 3-Dec 16.7 17-Jan 10.0 3-Mar 12.5 18-Apr 22.5 3-Jun 17.9 19-Jul 6.2 3-Sep 24.0 19-Oct 6.6 4-Dec 12.5 18-Jan 15.6 4-Mar 7.1 19-Apr 12.8 4-Jun 21.1 20-Jul 20.1 4-Sep 15.4 20-Oct 8.1 5-Dec 7.3 19-Jan 18.0 5-Mar 4.9 20-Apr 6.9 5-Jun 17.9 21-Jul 26.5 5-Sep 12.4 21-Oct 12.2 6-Dec 10.4 20-Jan 6.6 6-Mar 9.9 21-Apr 7.5 6-Jun 17.6 22-Jul 16.9 6-Sep 12.5 22-Oct 4.6 7-Dec 13.4 21-Jan 7.4 7-Mar 11.2 22-Apr 6.0 7-Jun 15.0 23-Jul 12.8 7-Sep 15.8 23-Oct 6.1 8-Dec 10.5 22-Jan 13.5 8-Mar 5.5 23-Apr 9.1 8-Jun 22.3 24-Jul 7.9 8-Sep 23.4 24-Oct 4.6 9-Dec 9.3 23-Jan 16.0 9-Mar 8.8 24-Apr 10.3 9-Jun 27.9 25-Jul 15.7 9-Sep 11.5 25-Oct 4.5 10-Dec 6.5 24-Jan 9.4 10-Mar 11.0 25-Apr 12.0 10-Jun 21.6 26-Jul 24.9 10-Sep 6.0 26-Oct 10.5 11-Dec 3.0 25-Jan 12.6 11-Mar 12.1 26-Apr 12.5 11-Jun 19.4 27-Jul 22.2 11-Sep 11.8 27-Oct 6.4 12-Dec 3.5 26-Jan 13.6 12-Mar 9.7 27-Apr 11.3 12-Jun 21.2 28-Jul 17.5 12-Sep 10.7 28-Oct 4.6 13-Dec 10.2 27-Jan 16.1 13-Mar 15.1 28-Apr 7.6 13-Jun 29.1 29-Jul 19.1 13-Sep 7.6 29-Oct 5.6 14-Dec 17.6 28-Jan 10.0 14-Mar 21.6 29-Apr 7.4 14-Jun 15.6 30-Jul 21.1 14-Sep 7.5 30-Oct 7.6 15-Dec 12.4 29-Jan 10.4 15-Mar 16.6 30-Apr 11.4 15-Jun 14.8 31-Jul 18.0 15-Sep 7.1 31-Oct 11.2 16-Dec 9.7 30-Jan 6.9 16-Mar 7.9 1-May 12.6 16-Jun 17.8 1-Aug 16.3 16-Sep 7.7 1-Nov 16.2 17-Dec 7.0 31-Jan 4.9 17-Mar 9.6 2-May 10.0 17-Jun 12.6 2-Aug 19.3 17-Sep 11.3 2-Nov 17.3 18-Dec 7.9 1-Feb 5.4 18-Mar 10.3 3-May 11.2 18-Jun 10.5 3-Aug 17.9 18-Sep 16.8 3-Nov 18.3 19-Dec 6.9 2-Feb 7.1 19-Mar 8.4 4-May 10.4 19-Jun 15.0 4-Aug 25.1 19-Sep 14.8 4-Nov 8.9 20-Dec 8.1 3-Feb 10.9 20-Mar 4.9 5-May 15.7 20-Jun 22.7 5-Aug 29.3 20-Sep 8.0 5-Nov 5.8 21-Dec 4.9 4-Feb 12.1 21-Mar 8.7 6-May 16.1 21-Jun 18.7 6-Aug 19.1 21-Sep 10.8 6-Nov 8.6 22-Dec 7.7 5-Feb 17.1 22-Mar 13.3 7-May 16.8 22-Jun 15.2 7-Aug 14.0 22-Sep 14.5 7-Nov 15.0 23-Dec 7.7 6-Feb 10.3 23-Mar 12.2 8-May 14.5 23-Jun 16.8 8-Aug 10.8 23-Sep 21.2 8-Nov 8.3 24-Dec 10.5 7-Feb 4.0 24-Mar 10.3 9-May 11.7 24-Jun 15.1 9-Aug 15.0 24-Sep 8.6 9-Nov 10.0 25-Dec 6.5 8-Feb 9.7 25-Mar 11.9 10-May 9.0 25-Jun 20.7 10-Aug 21.7 25-Sep 1.2 10-Nov 12.8 26-Dec 7.6 9-Feb 11.5 26-Mar 20.1 11-May 6.7 26-Jun 23.0 11-Aug 14.3 26-Sep 16.0 11-Nov 11.8 27-Dec 13.3 10-Feb 3.0 27-Mar 22.5 12-May 7.9 27-Jun 17.8 12-Aug 14.7 27-Sep 12.1 12-Nov 14.8 28-Dec 6.4 11-Feb 5.5 28-Mar 18.2 13-May 8.3 28-Jun 12.4 13-Aug 13.0 28-Sep 18.0 13-Nov 14.5 29-Dec 3.7 12-Feb 18.9 29-Mar 10.8 14-May 12.2 29-Jun 12.7 14-Aug 13.5 29-Sep 17.8 14-Nov 7.7 30-Dec 4.7 13-Feb 17.6 30-Mar 6.4 15-May 13.1 30-Jun 8.9 15-Aug 17.5 30-Sep 16.4 15-Nov 3.6 31-Dec 4.4 14-Feb 11.2 31-Mar 3.3 16-May 8.8 1-Jul 7.1 16-Aug 23.9 1-Oct 12.3 16-Nov 4.6 15-Feb 14.4 1-Apr 7.8 17-May 8.2 2-Jul 13.8 17-Aug 18.4 2-Oct 8.2 17-Nov 7.8

Annual 98th Percentile Concentration = 25.1 µg/m3 D-3

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

th Table E-2. 2008 Daily PM2.5 Concentrations Less Than or Equal to the 98 Percentile Date Conc. Date Conc. Date Conc. Date Conc. Date Conc. Date Conc. Date Conc. Date Conc. 1-Jan 10.4 16-Feb 15.1 2-Apr 10.5 18-May 11.1 3-Jul 17.1 18-Aug 18.7 3-Oct 12.3 18-Nov 4.4 2-Jan 5.4 17-Feb 11.8 3-Apr 8.2 19-May 7.7 4-Jul 19.8 19-Aug 21.5 4-Oct 19.5 19-Nov 8.2 3-Jan 10.0 18-Feb 3.4 4-Apr 9.7 20-May 13.6 5-Jul 14.3 20-Aug 20.1 5-Oct 23.7 20-Nov 11.1 4-Jan 16.4 19-Feb 4.5 5-Apr 6.9 21-May 12.1 6-Jul 11.5 21-Aug 18.4 6-Oct 19.8 21-Nov 5.3 5-Jan 11.2 20-Feb 4.8 6-Apr 6.3 22-May 10.0 7-Jul 14.3 22-Aug 16.7 7-Oct 21.7 22-Nov 8.9 6-Jan 11.1 21-Feb 11.9 7-Apr 7.9 23-May 13.3 8-Jul 12.2 23-Aug 13.8 8-Oct 12.2 23-Nov 14.0 7-Jan 10.2 22-Feb 20.1 8-Apr 9.8 24-May 11.2 9-Jul 11.1 24-Aug 19.0 9-Oct 5.1 24-Nov 12.7 8-Jan 11.4 23-Feb 11.4 9-Apr 16.5 25-May 17.7 10-Jul 9.7 25-Aug 17.6 10-Oct 10.2 25-Nov 9.7 9-Jan 8.1 24-Feb 19.3 10-Apr 13.3 26-May 14.2 11-Jul 16.4 26-Aug 15.4 11-Oct 10.7 26-Nov 12.8 10-Jan 9.4 25-Feb 18.2 11-Apr 11.0 27-May 15.4 12-Jul 21.5 27-Aug 12.6 12-Oct 5.6 27-Nov 16.6 11-Jan 5.7 26-Feb 12.8 12-Apr 8.8 28-May 13.9 13-Jul RC 28-Aug 12.1 13-Oct 5.9 28-Nov 17.2 12-Jan 8.9 27-Feb 5.5 13-Apr 6.3 29-May 9.3 14-Jul 11.7 29-Aug 10.1 14-Oct 9.7 29-Nov 16.6 13-Jan 18.1 28-Feb 9.7 14-Apr 5.1 30-May 14.5 15-Jul 18.9 30-Aug 17.2 15-Oct 12.8 30-Nov 4.5 14-Jan 11.0 29-Feb 12.1 15-Apr 7.9 31-May 20.5 16-Jul RC 31-Aug 19.9 16-Oct 16.4 1-Dec 7.5 15-Jan 11.8 1-Mar 9.6 16-Apr 8.2 1-Jun 15.3 17-Jul RC 1-Sep 19.4 17-Oct 12.0 2-Dec 10.6 16-Jan 10.7 2-Mar 5.6 17-Apr 14.7 2-Jun 11.5 18-Jul 12.8 2-Sep 18.2 18-Oct 7.9 3-Dec 16.7 17-Jan 10.0 3-Mar 12.5 18-Apr 22.5 3-Jun 17.9 19-Jul 6.2 3-Sep 24.0 19-Oct 6.6 4-Dec 12.5 18-Jan 15.6 4-Mar 7.1 19-Apr 12.8 4-Jun 21.1 20-Jul 20.1 4-Sep 15.4 20-Oct 8.1 5-Dec 7.3 19-Jan 18.0 5-Mar 4.9 20-Apr 6.9 5-Jun 17.9 21-Jul RC 5-Sep 12.4 21-Oct 12.2 6-Dec 10.4 20-Jan 6.6 6-Mar 9.9 21-Apr 7.5 6-Jun 17.6 22-Jul 16.9 6-Sep 12.5 22-Oct 4.6 7-Dec 13.4 21-Jan 7.4 7-Mar 11.2 22-Apr 6.0 7-Jun 15.0 23-Jul 12.8 7-Sep 15.8 23-Oct 6.1 8-Dec 10.5 22-Jan 13.5 8-Mar 5.5 23-Apr 9.1 8-Jun 22.3 24-Jul 7.9 8-Sep 23.4 24-Oct 4.6 9-Dec 9.3 23-Jan 16.0 9-Mar 8.8 24-Apr 10.3 9-Jun RC 25-Jul 15.7 9-Sep 11.5 25-Oct 4.5 10-Dec 6.5 24-Jan 9.4 10-Mar 11.0 25-Apr 12.0 10-Jun 21.6 26-Jul 24.9 10-Sep 6.0 26-Oct 10.5 11-Dec 3.0 25-Jan 12.6 11-Mar 12.1 26-Apr 12.5 11-Jun 19.4 27-Jul 22.2 11-Sep 11.8 27-Oct 6.4 12-Dec 3.5 26-Jan 13.6 12-Mar 9.7 27-Apr 11.3 12-Jun 21.2 28-Jul 17.5 12-Sep 10.7 28-Oct 4.6 13-Dec 10.2 27-Jan 16.1 13-Mar 15.1 28-Apr 7.6 13-Jun RC 29-Jul 19.1 13-Sep 7.6 29-Oct 5.6 14-Dec 17.6 28-Jan 10.0 14-Mar 21.6 29-Apr 7.4 14-Jun 15.6 30-Jul 21.1 14-Sep 7.5 30-Oct 7.6 15-Dec 12.4 29-Jan 10.4 15-Mar 16.6 30-Apr 11.4 15-Jun 14.8 31-Jul 18.0 15-Sep 7.1 31-Oct 11.2 16-Dec 9.7 30-Jan 6.9 16-Mar 7.9 1-May 12.6 16-Jun 17.8 1-Aug 16.3 16-Sep 7.7 1-Nov 16.2 17-Dec 7.0 31-Jan 4.9 17-Mar 9.6 2-May 10.0 17-Jun 12.6 2-Aug 19.3 17-Sep 11.3 2-Nov 17.3 18-Dec 7.9 1-Feb 5.4 18-Mar 10.3 3-May 11.2 18-Jun 10.5 3-Aug 17.9 18-Sep 16.8 3-Nov 18.3 19-Dec 6.9 2-Feb 7.1 19-Mar 8.4 4-May 10.4 19-Jun 15.0 4-Aug 25.1 19-Sep 14.8 4-Nov 8.9 20-Dec 8.1 3-Feb 10.9 20-Mar 4.9 5-May 15.7 20-Jun 22.7 5-Aug RC 20-Sep 8.0 5-Nov 5.8 21-Dec 4.9 4-Feb 12.1 21-Mar 8.7 6-May 16.1 21-Jun 18.7 6-Aug 19.1 21-Sep 10.8 6-Nov 8.6 22-Dec 7.7 5-Feb 17.1 22-Mar 13.3 7-May 16.8 22-Jun 15.2 7-Aug 14.0 22-Sep 14.5 7-Nov 15.0 23-Dec 7.7 6-Feb 10.3 23-Mar 12.2 8-May 14.5 23-Jun 16.8 8-Aug 10.8 23-Sep 21.2 8-Nov 8.3 24-Dec 10.5 7-Feb 4.0 24-Mar 10.3 9-May 11.7 24-Jun 15.1 9-Aug 15.0 24-Sep 8.6 9-Nov 10.0 25-Dec 6.5 8-Feb 9.7 25-Mar 11.9 10-May 9.0 25-Jun 20.7 10-Aug 21.7 25-Sep 1.2 10-Nov 12.8 26-Dec 7.6 9-Feb 11.5 26-Mar 20.1 11-May 6.7 26-Jun 23.0 11-Aug 14.3 26-Sep 16.0 11-Nov 11.8 27-Dec 13.3 10-Feb 3.0 27-Mar 22.5 12-May 7.9 27-Jun 17.8 12-Aug 14.7 27-Sep 12.1 12-Nov 14.8 28-Dec 6.4 11-Feb 5.5 28-Mar 18.2 13-May 8.3 28-Jun 12.4 13-Aug 13.0 28-Sep 18.0 13-Nov 14.5 29-Dec 3.7 12-Feb 18.9 29-Mar 10.8 14-May 12.2 29-Jun 12.7 14-Aug 13.5 29-Sep 17.8 14-Nov 7.7 30-Dec 4.7 13-Feb 17.6 30-Mar 6.4 15-May 13.1 30-Jun 8.9 15-Aug 17.5 30-Sep 16.4 15-Nov 3.6 31-Dec 4.4 14-Feb 11.2 31-Mar 3.3 16-May 8.8 1-Jul 7.1 16-Aug 23.9 1-Oct 12.3 16-Nov 4.6 15-Feb 14.4 1-Apr 7.8 17-May 8.2 2-Jul 13.8 17-Aug 18.4 2-Oct 8.2 17-Nov 7.8

Annual 98th Percentile Concentration = 25.1 µg/m3 RC = Above 98th Percentile and Removed from Consideration

D-4

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

th Table E-3. 2008 Daily PM2.5 Concentrations Less Than or Equal to the 98 Percentile by Quarter Season / Quarter 1 Season / Quarter 2 Season / Quarter 3 Season / Quarter 4 Date Conc. Date Conc. Date Conc. Date Conc. Date Conc. Date Conc. Date Conc. Date Conc. 1-Jan 10.4 16-Feb 15.1 1-Apr 7.8 17-May 8.2 1-Jul 7.1 16-Aug 23.9 1-Oct 12.3 16-Nov 4.6 2-Jan 5.4 17-Feb 11.8 2-Apr 10.5 18-May 11.1 2-Jul 13.8 17-Aug 18.4 2-Oct 8.2 17-Nov 7.8 3-Jan 10.0 18-Feb 3.4 3-Apr 8.2 19-May 7.7 3-Jul 17.1 18-Aug 18.7 3-Oct 12.3 18-Nov 4.4 4-Jan 16.4 19-Feb 4.5 4-Apr 9.7 20-May 13.6 4-Jul 19.8 19-Aug 21.5 4-Oct 19.5 19-Nov 8.2 5-Jan 11.2 20-Feb 4.8 5-Apr 6.9 21-May 12.1 5-Jul 14.3 20-Aug 20.1 5-Oct 23.7 20-Nov 11.1 6-Jan 11.1 21-Feb 11.9 6-Apr 6.3 22-May 10.0 6-Jul 11.5 21-Aug 18.4 6-Oct 19.8 21-Nov 5.3 7-Jan 10.2 22-Feb 20.1 7-Apr 7.9 23-May 13.3 7-Jul 14.3 22-Aug 16.7 7-Oct 21.7 22-Nov 8.9 8-Jan 11.4 23-Feb 11.4 8-Apr 9.8 24-May 11.2 8-Jul 12.2 23-Aug 13.8 8-Oct 12.2 23-Nov 14.0 9-Jan 8.1 24-Feb 19.3 9-Apr 16.5 25-May 17.7 9-Jul 11.1 24-Aug 19.0 9-Oct 5.1 24-Nov 12.7 10-Jan 9.4 25-Feb 18.2 10-Apr 13.3 26-May 14.2 10-Jul 9.7 25-Aug 17.6 10-Oct 10.2 25-Nov 9.7 11-Jan 5.7 26-Feb 12.8 11-Apr 11.0 27-May 15.4 11-Jul 16.4 26-Aug 15.4 11-Oct 10.7 26-Nov 12.8 12-Jan 8.9 27-Feb 5.5 12-Apr 8.8 28-May 13.9 12-Jul 21.5 27-Aug 12.6 12-Oct 5.6 27-Nov 16.6 13-Jan 18.1 28-Feb 9.7 13-Apr 6.3 29-May 9.3 13-Jul RC 28-Aug 12.1 13-Oct 5.9 28-Nov 17.2 14-Jan 11.0 29-Feb 12.1 14-Apr 5.1 30-May 14.5 14-Jul 11.7 29-Aug 10.1 14-Oct 9.7 29-Nov 16.6 15-Jan 11.8 1-Mar 9.6 15-Apr 7.9 31-May 20.5 15-Jul 18.9 30-Aug 17.2 15-Oct 12.8 30-Nov 4.5 16-Jan 10.7 2-Mar 5.6 16-Apr 8.2 1-Jun 15.3 16-Jul RC 31-Aug 19.9 16-Oct 16.4 1-Dec 7.5 17-Jan 10.0 3-Mar 12.5 17-Apr 14.7 2-Jun 11.5 17-Jul RC 1-Sep 19.4 17-Oct 12.0 2-Dec 10.6 18-Jan 15.6 4-Mar 7.1 18-Apr 22.5 3-Jun 17.9 18-Jul 12.8 2-Sep 18.2 18-Oct 7.9 3-Dec 16.7 19-Jan 18.0 5-Mar 4.9 19-Apr 12.8 4-Jun 21.1 19-Jul 6.2 3-Sep 24.0 19-Oct 6.6 4-Dec 12.5 20-Jan 6.6 6-Mar 9.9 20-Apr 6.9 5-Jun 17.9 20-Jul 20.1 4-Sep 15.4 20-Oct 8.1 5-Dec 7.3 21-Jan 7.4 7-Mar 11.2 21-Apr 7.5 6-Jun 17.6 21-Jul RC 5-Sep 12.4 21-Oct 12.2 6-Dec 10.4 22-Jan 13.5 8-Mar 5.5 22-Apr 6.0 7-Jun 15.0 22-Jul 16.9 6-Sep 12.5 22-Oct 4.6 7-Dec 13.4 23-Jan 16.0 9-Mar 8.8 23-Apr 9.1 8-Jun 22.3 23-Jul 12.8 7-Sep 15.8 23-Oct 6.1 8-Dec 10.5 24-Jan 9.4 10-Mar 11.0 24-Apr 10.3 9-Jun RC 24-Jul 7.9 8-Sep 23.4 24-Oct 4.6 9-Dec 9.3 25-Jan 12.6 11-Mar 12.1 25-Apr 12.0 10-Jun 21.6 25-Jul 15.7 9-Sep 11.5 25-Oct 4.5 10-Dec 6.5 26-Jan 13.6 12-Mar 9.7 26-Apr 12.5 11-Jun 19.4 26-Jul 24.9 10-Sep 6.0 26-Oct 10.5 11-Dec 3.0 27-Jan 16.1 13-Mar 15.1 27-Apr 11.3 12-Jun 21.2 27-Jul 22.2 11-Sep 11.8 27-Oct 6.4 12-Dec 3.5 28-Jan 10.0 14-Mar 21.6 28-Apr 7.6 13-Jun RC 28-Jul 17.5 12-Sep 10.7 28-Oct 4.6 13-Dec 10.2 29-Jan 10.4 15-Mar 16.6 29-Apr 7.4 14-Jun 15.6 29-Jul 19.1 13-Sep 7.6 29-Oct 5.6 14-Dec 17.6 30-Jan 6.9 16-Mar 7.9 30-Apr 11.4 15-Jun 14.8 30-Jul 21.1 14-Sep 7.5 30-Oct 7.6 15-Dec 12.4 31-Jan 4.9 17-Mar 9.6 1-May 12.6 16-Jun 17.8 31-Jul 18.0 15-Sep 7.1 31-Oct 11.2 16-Dec 9.7 1-Feb 5.4 18-Mar 10.3 2-May 10.0 17-Jun 12.6 1-Aug 16.3 16-Sep 7.7 1-Nov 16.2 17-Dec 7.0 2-Feb 7.1 19-Mar 8.4 3-May 11.2 18-Jun 10.5 2-Aug 19.3 17-Sep 11.3 2-Nov 17.3 18-Dec 7.9 3-Feb 10.9 20-Mar 4.9 4-May 10.4 19-Jun 15.0 3-Aug 17.9 18-Sep 16.8 3-Nov 18.3 19-Dec 6.9 4-Feb 12.1 21-Mar 8.7 5-May 15.7 20-Jun 22.7 4-Aug 25.1 19-Sep 14.8 4-Nov 8.9 20-Dec 8.1 5-Feb 17.1 22-Mar 13.3 6-May 16.1 21-Jun 18.7 5-Aug RC 20-Sep 8.0 5-Nov 5.8 21-Dec 4.9 6-Feb 10.3 23-Mar 12.2 7-May 16.8 22-Jun 15.2 6-Aug 19.1 21-Sep 10.8 6-Nov 8.6 22-Dec 7.7 7-Feb 4.0 24-Mar 10.3 8-May 14.5 23-Jun 16.8 7-Aug 14.0 22-Sep 14.5 7-Nov 15.0 23-Dec 7.7 8-Feb 9.7 25-Mar 11.9 9-May 11.7 24-Jun 15.1 8-Aug 10.8 23-Sep 21.2 8-Nov 8.3 24-Dec 10.5 9-Feb 11.5 26-Mar 20.1 10-May 9.0 25-Jun 20.7 9-Aug 15.0 24-Sep 8.6 9-Nov 10.0 25-Dec 6.5 10-Feb 3.0 27-Mar 22.5 11-May 6.7 26-Jun 23.0 10-Aug 21.7 25-Sep 1.2 10-Nov 12.8 26-Dec 7.6 11-Feb 5.5 28-Mar 18.2 12-May 7.9 27-Jun 17.8 11-Aug 14.3 26-Sep 16.0 11-Nov 11.8 27-Dec 13.3 12-Feb 18.9 29-Mar 10.8 13-May 8.3 28-Jun 12.4 12-Aug 14.7 27-Sep 12.1 12-Nov 14.8 28-Dec 6.4 13-Feb 17.6 30-Mar 6.4 14-May 12.2 29-Jun 12.7 13-Aug 13.0 28-Sep 18.0 13-Nov 14.5 29-Dec 3.7 14-Feb 11.2 31-Mar 3.3 15-May 13.1 30-Jun 8.9 14-Aug 13.5 29-Sep 17.8 14-Nov 7.7 30-Dec 4.7 15-Feb 14.4 16-May 8.8 15-Aug 17.5 30-Sep 16.4 15-Nov 3.6 31-Dec 4.4 Seasonal / Quarterly Maximum 22.5 Seasonal / Quarterly Maximum 23.0 Seasonal / Quarterly Maximum 25.1 Seasonal / Quarterly Maximum 23.7

Seasonal / Quarterly Maximum Concentration RC = Above 98th Percentile and Removed from Consideration

D-5

Does not represent final Agency action; Draft for public review and comment; 02/10/2020

Table E-4. Resulting Average of Seasonal (or Quarterly) Maximums for Inclusion into AERMOD

Seasonal / Quarterly Average Highest Monitored Concentration (From Annual Datasets Equal To and Less Than the 98th Percentile) Q1 Q2 Q3 Q4 2008 22.5 23.0 25.1 23.7 2009 21.1 20.7 21.2 19.8 2010 20.7 22.6 23.5 20.7 Average 21.433 22.100 23.267 21.400

(Note, the complete datasets for 2009 and 2010 are not shown in Appendix D but would follow the same steps as for 2008)

D-6

United States Office of Air Quality Planning and Standards Publication No. EPA-457/P-20-002 Environmental Protection Air Quality Assessment Division February 2020 Agency Research Triangle Park, NC